Create a risk map

A comprehensive plan to address homelessness includes programs to prevent homelessness before it starts and programs to support people already experiencing homelessness. This lesson focuses on prevention. It answers the question, Where are people in Los Angeles County becoming homeless?

Prevention-based interventions are more cost-effective than the amount to run supportive services for those experiencing chronic homelessness. To prevent homelessness, however, you must know where it begins. The point-in-time count data for Los Angeles County indicates where homeless people are, not where they were when they became homeless. You'll predict where people are becoming homeless based on key factors known to contribute to homelessness. These factors include poverty, addiction, lack of affordable housing, unemployment, mental illness, domestic violence, and high health care costs. For this analysis, you'll give all factors equal weighting.

Open the project

First, you'll download and open an ArcGIS Pro project with relevant demographic data.

  1. Download the homeless-data compressed folder.
  2. Right-click the downloaded folder and extract it to a location you can easily find, such as your Documents folder.
  3. Open the homeless-data folder.

    The folder contains a file geodatabase (.gdb) with data, an index file, an ArcGIS Pro project file (.aprx), and an ArcGIS toolbox (.tbx).

  4. If you have ArcGIS Pro installed on your machine, double-click the Homelessness project file (it may have an .aprx extension). If prompted, sign in using your licensed ArcGIS Online or ArcGIS Enterprise account.

    Note:
    If you don't have ArcGIS Pro or an ArcGIS account, you can sign up for an ArcGIS free trial.

    The project contains a map of Los Angeles County.

    The project opens showing Los Angeles County in California.

    In addition to the basemap, the project contains four layers, one of which is the county outline. The other three layers contain the county's census tracts with different types of demographic information as attributes.

    Next, you will explore the Risk Factors layer.

  5. In the Contents pane, right-click Risk Factors and choose Attribute Table.

    Attribute Table option

    The table appears. The table contains demographic information about income, mental illness, substance abuse, and other factors that can contribute to people becoming homeless for each census tract. This data was obtained using the Enrich tool and was converted into rates where appropriate.

    Note:

    Learn more about the Enrich tool.

    Post-traumatic stress disorder (PTSD) among veterans was estimated using American Community Survey (ACS) veteran data and PTSD prevalence data. Prevalence of severe mental illness was computed using Esri demographic data and mental illness prevalence data. Domestic violence incidents were extracted from the Los Angeles County Sheriff's Department crime incident data. Data indicating change (in unemployment rate, for example) was obtained from past ACS data sources.

  6. Close the table.

Copy a census tract

As you saw in the table, there are many risk factors for homelessness. You could symbolize the layer to show an individual risk factor, but to predict where people are becoming homeless, you must consider all risk factors at once.

To make your risk map, you'll create a hypothetical census tract that has the worst values found across the county for each risk factor. Then, you'll rank all census tracts by how similar they are to the worst-case tract. Doing so will give you a picture of homelessness risk across the county.

First, you'll copy an existing census tract to be the base for your hypothetical tract.

  1. In the Contents pane, uncheck the County Outline layer.

    With the outline turned off, it'll be easier to select features in the Risk Factors layer.

  2. On the ribbon, click the Map tab. In the Selection group, click the Select button.

    Select button

  3. On the map, click any census tract to select it.

    Selected census tract

  4. In the Contents pane, right-click Risk Factors. Point to Data and choose Export Features.

    The Export Features pane appears. The tool creates a copy of a feature layer. If one or more features of the layer are selected, only the selected features will be copied.

  5. For Input Features, confirm that Risk Factors is chosen. For Output Feature Class, change the text to Target_Tract.
    Note:

    The tool indicates how many selected features will be processed.

    Export Features tool parameters

  6. Click OK.

    The Target_Tract layer is added to the map and the Contents pane. The copied tract overlaps with the original tract, so you can't see it. You'll edit the copied tract to move it outside of the study area.

  7. With the Select tool, click anywhere on the basemap to clear the selection.
  8. In the Contents pane, turn off the Risk Factors layer. Use the Select tool to select the copied census tract.
  9. On the ribbon, click the Edit tab. In the Tools group, click Move.

    Move button

  10. Drag the selected census tract outside of Los Angeles County.

    Move the tract outside of Los Angeles County.
    Tip:

    The basemap contains a faint outline of the county. If you have difficulty seeing the outline, you can turn the County Outline layer back on.

  11. At the bottom of the map, click Finish.

    Finish button
    Tip:

    Alternatively, you can press F2 to finish editing.

  12. On the ribbon, in the Manage Edits group, click Save.

    Save button

    The Save Edits window appears, asking you to confirm that you want to save your edits.

  13. In the Save Edits window, click Yes.
  14. Close the Modify Features pane.
  15. In the Contents pane, turn on the County Outline and Risk Factors layers. On the ribbon, in the Selection group, click Clear.

    Clear button

    The selected feature is cleared. Your map shows the original census tracts and your new target tract.

    Project map with copied census tract

    Note:

    It's possible that the Move editing tool is still active. If you want to navigate the map, click the Map tab. In the Navigate group, click Explore.

  16. On the Quick Access Toolbar, click the Save button to save the project.

    Save button

  17. If you receive a message about ArcGIS Pro versions, click Yes.

Edit attribute data

Next, you'll edit the attribute data of the target tract so that it contains the worst values countywide for each homelessness risk factor.

  1. Open the attribute tables for the Target_Tract and Risk Factors layers.

    Both layers contain a 2017 Total Population field. While this field gives useful context, it isn't a risk factor. You'll delete it from the table.

  2. In the Target_Tract table, right-click the 2017 Total Population field (near the end of the table) and choose Delete.

    Delete the 2017 Total Population field.

  3. In the Delete Field window, click Yes.

    The field is deleted. Next, you'll edit the remaining fields. Currently, you can only view one table at a time. You'll need to compare the tables frequently, so you'll rearrange them to view both at once.

  4. Drag the Risk Factors table tab to reposition it above the Target_Tract table.

    Moving a table

    The tables are displayed one above the other.

    Tip:

    You can rearrange either table in any position you prefer. The objective is to view both tables at once.

  5. In the Risk Factors table, right-click the 2011-2015 ACS Households with Income Below Poverty Level field and choose Sort Descending.

    Sort Descending option

    Tip:

    Alternatively, double-click the field name to sort from smallest to largest. Double-click a second time to sort from largest to smallest.

    The field is sorted from the largest to smallest value. The highest number of households with income below the poverty level in a single census tract is 1,809. This is the worst value in the county, so you'll edit the Target_Tract table accordingly.

  6. In the Target_Tract table, double-click the value for the TRACT field and type 0. In the same way, edit the value for the 2011-2015 ACS Households with Income Below Poverty Level field and type 1809.

    Editing table values

    Normally, you would repeat the process for all other risk factors. For the purposes of this exercise, you'll be provided the remaining worst-case values.

  7. Edit the Target_Tract table with the following values:
    Tip:

    Press the Tab key to quickly switch to the next field.

    Field nameValue

    2011-2015 ACS Percent of Households with Income Below Poverty Level

    100

    Change in number of HH below poverty level 13-17

    311

    ACS HHs w/Public Assist Income

    525

    ACS % HHs w Public Assistance Income

    42.40711

    Change in number of HH w Public Assistance 13-17

    145

    ACS HHs w/1+ Persons w/Disability

    1068

    ACS % HHs w 1+ Persons w/Disability

    100

    ACS HHs/Gross Rent 50+% of Income

    1132

    % HHs paying 50+% of income for rent

    100

    Change in number of HHs paying 50+% of income for rent

    207

    2017 Unemployed Population 16+

    371

    2017 Unemployment Rate

    52

    Change in Unemployment Rate 10-17

    52

    No Health Insurance Pop 18+

    2760

    % No Health Insurance Pop 18+

    74.94635

    % Veterans with PTSD

    25

    Est Veterans w PTSD

    92

    Est Serious Mental Illness Cnt

    1004

    Est % w a Serious Mental Illness

    12.5

    Substance Abuse Incidents

    233

    Substance Abuse Rate

    600

    Domestic Violence Incidents

    91

    Domestic Violence Rate

    4.545455

    Note:

    Be sure to press Enter after changing the final value (otherwise you can't save your changes).

    Your target tract now has the worst-case values countywide for each risk factor.

  8. On the ribbon, click the Edit tab. Click Save and in the Save Edits window, click Yes to save all edits.
  9. Close the Risk Factors and Target_Tract tables.
  10. Save the project.

Rank tracts by similarity to the target

Next, you'll run the Similarity Search analysis tool to compare all census tracts to your target tract. The tool ranks tracts by how similar they are based on their attributes. The tracts with the smallest ranking will be those with the highest risk for homelessness.

  1. On the ribbon, click the Analysis tab. In the Geoprocessing group, click Tools.

    Tools button

  2. In the Geoprocessing pane, search for Similarity Search. Click the Similarity Search tool.

    Similarity Search tool

    The tool's parameters appear. First, you'll choose the feature layers to compare.

    Note:

    Learn more about the Similarity Search tool.

  3. In the Similarity Search tool pane, for Input Features to Match, choose Target_Tract. For Candidate Features, choose Risk Factors.
  4. For Output Features, change the output name to Risk_Surface.

    Similarity Search tool input and output features parameters

    If checked, the Collapse Output To Points parameter creates a point feature layer instead of matching the existing geometry. You'll leave this unchecked so that your output retains the shape of the tracts.

    Note:

    The Collapse Output to Points parameter is only available in ArcGIS Pro Advanced. Learn more about the Similarity Search tool.

    The next two parameters determine how features will be ranked. You want to rank tracts that are most similar to the target tract feature based on attribute values.

  5. For Most Or Least Similar, confirm that Most similar is chosen. For Match Method, confirm that Attribute values is chosen.

    The next parameter, Number Of Results, determines how many features are ranked. If you type 0, all features will be ranked.

  6. For Number Of Results, type 0.

    Similarity Search tool matching parameters

  7. For Attributes Of Interest, click Select All. Uncheck UniqID, Shape_Length, and Shape_Area.

    Similarity Search tool attribute parameters

    These three attributes aren't risk factors, so they shouldn't be part of the comparison.

    Next, you'll append two fields from the Risk Factors layer onto the output layer. These fields will contain the name of each tract and its total population. You previously deleted the population field from the Target_Tract layer because it wasn't a risk factor, but you'll want to add it again after the risk factors are compared.

  8. Expand Additional Options. For Fields To Append To Output, check TRACT and 2017 Total Population.

    Similarity Search tool additional parameters

  9. Click Run.

    The tool runs and the output is added to the map.

    Similarity Search output layer

    The darker blue tracts are those that are most similar to the target tract. Many areas with high risk for homelessness are located in the southern part of the county, which isn't too surprising because it coincides with the most urban areas of Los Angeles. However, there are also many high-risk tracts in the northern desert region, near Lancaster and Palmdale.

  10. In the Contents pane, turn off the Target_Tract and Risk Factors layers. Drag the County Outline layer above the Risk_Surface layer.
  11. Save the project.

Symbolize the tracts

The default symbology is okay, but the larger census tracts in the north are emphasized more than smaller ones in the south. It would be better if tracts with higher population were emphasized instead.

Symbolizing each tract with a point feature that represents the tract's population might work, but in Los Angeles that would create many overlapping circles. Instead, you'll de-emphasize census tracts with small populations by increasing their transparency.

  1. In the Contents pane, right-click the Risk_Surface layer and choose Symbology.

    Symbology option

    Note:

    You may receive a warning about an exclusion clause. The exclusion clause is for the target tract, which you'll later remove. You can ignore the warning.

    First, you'll change the color scheme so there is a starker difference between the most and least similar values.

  2. In the Symbology pane, for Color scheme, click the down arrow and check Show names. Choose Purple-Red (Continuous).

    Purple-Red (Continuous) color scheme and Show names checked

    The color scheme is applied automatically. By default, the least similar features have the darkest colors, which isn't what you want.

  3. On the Classes tab, click More and choose Reverse symbol order.

    Reverse symbol order

    Tracts with higher risk factor values for becoming homeless have darker colors, while tracts with fewer risk factors have a lighter color.

    You'll also remove the target tract, which no longer needs to appear on the map.

  4. Click More again and uncheck Show excluded values.

    The target tract no longer appears on the map. Another useful change to improve the readability of a map with many small features is to reduce the outline width.

  5. Click More and choose Format all symbols.
  6. Click the Properties tab. Change Outline width to 0.25 pt.

    Outline width parameter

  7. Click Apply.

    The thinner outlines make the smaller features easier to see. Next, you'll adjust the transparency of tracts with low population to de-emphasize them.

  8. In the Symbology pane, click the button for more options and choose Vary symbology by attribute.

    Vary symbology by attribute option

  9. Expand Transparency. For Field, choose 2017 Total Population.

    Transparency field

    Transparency is automatically applied. Tracts with high population values have 30 percent transparency by default, while those with low population have 70 percent. It would be better if tracts with high populations had no transparency, so they stand out more.

  10. Under Transparency range, change High values to 0%.

    High values parameter

    The range of transparency is still pretty broad, and only the highest population value has 0 percent transparency. You'll adjust the range so all tracts with a population higher than 8,000 have 0 percent transparency, while all tracts with a population less than 4,000 have 70 percent transparency. This way, differences between high and low values will appear more clearly.

  11. On the graph of values, double-click the upper marker on 0, type 4000, and press Enter. Double-click the other marker on 14595 and type 8000, and press Enter.

    Graph of values for transparency

    The graph includes a histogram of the values, so you have an idea of how many tracts will have more or less transparency. The transparency is applied automatically.

    Symbolized risk map

    Now, the worst-case tracts with the largest population are emphasized. The map gives a better idea of the most important areas to target homelessness prevention.

  12. Close the Symbology pane and save the project.

You have created a map of census tracts in Los Angeles County with the highest risk for generating new homelessness. You then symbolized the map to account for tracts with more residents.

Note:

This workflow can be used to create any kind of risk map. Start by identifying the administrative boundaries you want to analyze (ZIP Codes, census tracts, or neighborhoods, for example) and gather available data for that area that reflects risk. The Enrich tool provides data for any administrative boundaries you choose. If some of your data is only available for other administrative boundaries, consider using the sample Apportion Geoprocessing Tool or the Summarize Within tool to distribute the data from one geometry to another (for example, ZIP Codes to neighborhoods). If you have point data (for example, crime incidents), use the Spatial Join tool to count the number of points within each polygon. If some of the risk factor data isn't available, consider estimating it (see Mental Illness and PTSD, for example).

Don't worry if your data is incomplete; you can always repeat the analysis when additional variables become available. Move forward with what you have. The risk maps you obtain, even with partial data, will still be valuable.

Next, you'll identify the best locations for targeted prevention programs.


Prioritize prevention

Previously, you analyzed all risk factors associated with homelessness to find the locations at highest risk for generating new homelessness. Next, you'll analyze the spatial patterns of specific risk factors. Doing so will allow you to implement the right prevention programs in the right places, helping to ensure the best outcomes and the largest impacts.

You will aim to answer the question: Where will targeted prevention programs have their best chance of making a difference?

For the purposes of this exercise, you'll assume that Los Angeles County has decided to focus initial prevention efforts in two areas: unemployment and the lack of affordable housing.

Find top prevention program candidates

You'll use the same target tract you created earlier in this lesson. First, you'll rank tracts based on the unemployment risk factors. Then, you'll rank them based on the lack of affordable housing risk factors.

  1. If necessary, open the Homelessness project in ArcGIS Pro.
  2. Turn off all layers except the County Outline layer and the basemap.
  3. In the Geoprocessing pane, search for and open the Similarity Search tool.
    Tip:

    To open the Geoprocessing pane, on the Analysis tab, in the Geoprocessing group, click Tools.

  4. In the Similarity Search tool pane, set the following parameters:
    • For Input Features To Match, choose Target_Tract.
    • For Candidate Features, choose Risk Factors.
    • For Output Features, type Unemployment_Project.
    • Ensure Collapse Output To Points is unchecked.
    • Confirm Most Or Least Similar is set to Most Similar.
    • Confirmed Match Method is set to Attribute values.
    • For Number Of Results, type 25.

    Similarity Search tool parameters

  5. Continue setting the following parameters:
    • For Attributes Of Interest, check 2017 Unemployment Population 16+, 2017 Unemployment Rate, and Change in Unemployment Rate 10-17.
    • Expand Additional Options and under Fields To Append To Output, check TRACT.

    Similarity Search attribute and additional parameters

  6. Click Run.

    The tool runs and the output is added to the map. Twenty-five tracts with attributes similar to the worst-case target tract for unemployment risk factors are shown. The dark blue tracts are most similar, meaning they have the most severe unemployment risk factors. Later, you'll improve the symbology.

    Priority locations for prevention programs addressing unemployment

    Access to affordable housing also plays an important role in the dynamics of homelessness, and not just in Los Angeles. You'll use the Similarity Search tool again. This time, you'll identify priority locations for programs aimed at preventing homelessness by focusing on the lack of affordable housing.

  7. If necessary, open the Similarity Search tool.

    You'll use risk factor variables for households paying more than 50 percent of their total income for rent. It's possible that people choose rents they cannot afford, but it's more likely that they are paying more than 50 percent of their incomes for rent because affordable housing isn't available.

  8. In the Similarity Search tool, uncheck the three unemployment variables and check the following:
    • ACS HHs/Gross Rent 50+ % of Income
    • % HHs paying 50+ % of income for rent
    • Change in number of HHs paying 50+ % of Income for rent
  9. Ensure Collapse Output To Points is unchecked.
  10. Change Output Features to Affordable_Housing_Project.
  11. Click Run.

    The tool runs and the output is added to the map and Contents pane.

  12. Turn off all layers except the Affordable_Housing_Project layer and the basemap.

    The Affordable_Housing_Project layer shows 25 census tracts with the highest risk values for housing-related risk factors.

    Priority locations for prevention programs designed to address the lack of affordable housing

Symbolize program candidates

Both the Unemployment_Project and Affordable_Housing_Project layers have 26 features, representing the 25 worst-case tracts for each project and the target tract. First, you'll delete the target tract record from each output so it doesn't appear in your final map. Then, you'll symbolize both layers using graduated symbols so the priority tracts are easier to see.

  1. In the Contents pane, right-click the Unemployment_Project layer and choose Attribute Table.
  2. Select the target tract record (the first record in the table with a MATCH_ID value of 1).

    Target tract is selected

  3. On the ribbon, click the Edit tab. In the Features group, click Delete. In the Delete window that appears, click Yes.
  4. In the Manage Edits group, click Save and in the Save Edits window, click Yes.
  5. Use what you have learned to delete the target tract feature from the Affordable_Housing_Project layer.

    Tip:

    If the target tract still appears, click the Refresh button on the map.

    The default symbology makes it difficult to see the smallest candidate project tracts. To improve the symbology, you'll use graduated symbols to reflect both location and the number of people, or number of households, potentially impacted if a prevention program is initiated. You'll create three categories (high, medium, and low) with an equal number of tracts in each category.

  6. Close the tables.
  7. In the Contents pane, turn on the Unemployment_Project layer. Right-click the layer and choose Symbology.
  8. In the Symbology pane, change the following parameters:
    • For Primary symbology, choose Graduated Symbols.
    • For Field, choose 2017 Unemployment Population 16+.
    • For Method, choose Quantile.
    • For Classes, choose 3.
    • For Minimum size, choose 5 pt.
    • For Maximum size, choose 20 pt.

    Symbology parameters for the Unemployment_Project layer

  9. Click the Background symbol.

    Background symbol

  10. On the Properties tab, click the Layers button.

    Layers button

  11. For Color, choose No color.

    Setting the background color to No color

  12. On the Symbology pane, click Apply.
  13. To return to the primary Symbology pane, click the Back button.

    Back button

  14. Click the Template symbol.

    Modifying the Template symbol

    You'll also add a gradient effect to give the features a 3D appearance.

  15. Change Shape fill symbol to Buffered gradient fill. For Colors, choose Malachite Green and Indicolite Green. For Outline color, choose Gray 40%.
  16. Change Outline width to 0.5 pt. Click Apply.

    Symbol with a 3D effect

  17. Click the Back button. Set the class breaks (Upper value) and Label to the following values:
    • For the smallest symbol, change Upper value to 99 and Label to Fewer than 100.
    • For the middle symbol, change Upper value to 200 and Label to 100 to 200.
    • For the largest symbol, change Upper value to 371 and Label to More than 200.

    Class breaks and labels

  18. In the Contents pane, drag the County Outline layer to the top of the list of layers and if necessary, turn it on.

    The map displays the gradient symbols.

    Priority locations for programs addressing unemployment

  19. Follow the same process to symbolize the Affordable_Housing_Project layer using the ACS HHs/Gross Rent 50+ % of Income field with similar symbols as the previous layer, changing only the following parameters:
    • For Colors, choose Dark Amethyst and Lepidolite Lilac.
    • For the smallest symbol, change Upper value to 499 and Label to Fewer than 500.
    • For the middle symbol, change Upper value to 900 and Label to 500 to 900.
    • For the largest symbol, change Upper value to 1023 and Label to More than 900.

    Symbology for affordable housing priority map

    The symbology is applied to the map.

    Priority locations for programs addressing the lack of affordable housing

  20. Close the Symbology pane. Save the project.

You have identified targeted opportunities for preventing people from becoming homeless. The prevention programs you considered were for unemployment and the lack of affordable housing, but the same workflow could be used to identify priority locations for any risk factor.

Next, you'll use maps and charts to explore the characteristics of people who are currently experiencing homelessness.


Map homeless communities

Previously, you determined where targeted prevention programs would have the best chance of making a difference. In general, you were concerned with preventing homelessness before it starts. The remaining lessons focus on providing support and aid to people already experiencing homelessness.

Next, you'll use maps and charts to reveal differences among homeless communities. You'll answer several questions: Who are the homeless? What are their needs? Where are they located?

Find concentrations of homeless people

First, you'll perform hot spot analysis on data from the point-in-time count conducted in January 2017 for the Los Angeles Continuum of Care. The Continuum of Care region encompasses most of the county, but excludes the cities of Pasadena, Glendale, and Long Beach. Your analysis will pinpoint statistically significant concentrations of people experiencing homelessness. The point-in-time count specifies if a person experiencing homelessness on the night of the count was living in a shelter for the homeless (sheltered) or if they were living in a tent, encampment, or on the streets (unsheltered). You will analyze the concentration of people who were unsheltered.

  1. If necessary, open the Homelessness project in ArcGIS Pro.
  2. Turn off all layers except the County Outline layer and the basemap.
  3. Open the Geoprocessing pane. Search for and open the Optimized Hot Spot Analysis tool.
    Note:

    Learn more about the Optimized Hot Spot Analysis tool.

  4. For Input Features, choose the Homeless Population Data layer. For Output Features, change the output name to Unsheltered_Density_Hot_Spots.

    Large tracts often have more people (homeless or otherwise) because they cover large areas. If you were to analyze the number of unsheltered homeless people in each tract, the largest tracts would likely stand out. Rather than tract size, however, you're interested in understanding where unsheltered homeless people concentrate—where they cluster spatially. So instead of counts, you'll use a density variable calculated by dividing the number of unsheltered people in each tract by the tract's square area. Density values have already been calculated for you.

  5. For the Analysis Field parameter, choose Unsheltered Homeless Pop Density.

    The tool doesn't require a Distance Band value, but providing one ensures your hot spot maps are created using the same scale of analysis. This is important when you want to visually compare two or more hot spot maps.

  6. Expand Override Settings. For the Distance Band parameter, type 20000 and choose US Survey Feet from the other drop-down menu.
    Tip:

    When you run the Optimized Hot Spot Analysis tool without a distance band, the tool examines the structure of your data and suggests an appropriate distance band. The value of 20,000 feet was obtained by rounding down the distance suggested by the tool when no distance band was given.

    Optimized Hot Spot Analysis parameters

  7. Click Run.

    The tool runs and the output appears on the map.

    Resulting hot spot analysis map

    You'll improve the symbology by making the polygon outlines thinner.

  8. In the Contents pane, right-click the Unsheltered_Density_Hot_Spots layer and choose Symbology.
  9. In the Symbology pane, on the Classes tab, click More and choose Format all symbols.
  10. On the Properties tab, set Outline width to 0.2 and click Apply.
  11. In the Contents pane, drag the County Outline layer above the hot spot layer.

    The red areas of the hot spot layer are areas with statistically significant concentrations of people experiencing homelessness. The map shows there is a red area in the center of the county, which includes downtown Los Angeles where Skid Row is located and another major hot spot on the west side of the county around Venice Beach, a popular tourist location, and Santa Monica, where a major Veterans Affairs healthcare facility is located.

    Concentrations of unsheltered homeless densities

    The hot spot statistic used by the Optimized Hot Spot Analysis tool compares the actual distribution of high- and low-density values to a hypothetical random distribution. Red areas are locations where the spatial clustering of high densities deviates strongly and significantly from a random pattern. Similarly, the blue areas are locations where clustering of low densities is too intense to reflect a random pattern.

  12. If necessary, reopen the Optimized Hot Spot Analysis tool.
  13. For Output Features, type Sheletered_Density_Hot_Spots. For Analysis Field, choose Sheltered Homeless Pop Density.

    Optimized Hot Spot Analysis parameters

  14. Click Run.

    The tool runs and the output is added to the map. Both sheltered and unsheltered homeless densities are most concentrated in and around downtown Los Angeles.

    Note:

    To learn more about the history of homelessness in Los Angeles, read Combating Homelessness in Los Angeles County and Understanding LA's Homelessness Issues.

    Concentrations of sheltered homeless densities

  15. Save the project.

Chart homeless community characteristics

Knowing where homeless people are concentrating is an important first step, but it must be followed by analyses to understand differences among communities of people experiencing homelessness. Doing so leads to better decision making about the type of support and resources needed in each location.

The people experiencing homelessness in Venice Beach or Hollywood, for example, have different characteristics (and consequently different needs) than the people experiencing homelessness in Skid Row. While Skid Row has the smallest physical area, it contains the largest number of people experiencing homelessness and 57 percent have access to emergency shelters, transitional housing, or safe havens. In Venice Beach, on the other hand, only 12 percent of the people experiencing homelessness have access to shelters. The risk factors contributing to homelessness are also different among homeless communities. You'll create charts to explore these differences.

  1. Open the table for the Homeless Population Data layer.

    The homeless characteristics data (risk factors, race/ethnicity, gender, and age) have been apportioned from the City of Los Angeles Council District reports as well as from individual reports for Hollywood, Skid Row, and Venice Beach. Tracts outside the Council District areas don't have values (the values appear as Null). You'll select and then analyze the apportioned data.

  2. In the Geoprocessing pane, search for and open the Select Layer By Attribute tool. For Input Rows, choose Homeless Population Data. For Selection Type, ensure that New selection is chosen.
  3. Under Expression, construct the clause Where Homeless Veterans is not null.
  4. Click Add Clause. Construct the clause And Total Homeless People is greater than or equal to 10.

    Select Layer By Attribute parameters

  5. Click Run.

    Census tracts within the Council District areas, where the Homeless Veterans field isn't Null, and with at least 10 people experiencing homelessness are selected. Next, you'll create a feature class of the selected tracts.

  6. In the Geoprocessing pane, search for and open the Copy Features tool.
    Note:

    Learn more about the Copy Features tool.

  7. In the Copy Features tool pane, for Input Features, choose Homeless Population Data. For Output Feature Class, type Homeless_Tapestry.
  8. Click Run.

    The tool runs and the output layer is added to the map and to the Contents pane. The original layer still has selected features.

  9. On the ribbon, on the Map tab, in the Selection group, click Clear.

    Clear button

  10. Close the table.
  11. In the Contents pane, drag the County Outline layer to the top of the list of layers and turn off all layers except the Homeless_Tapestry, County Outline, and basemap layers.

    The Homeless_Tapestry layer shows the census tracts where the total population of people experiencing homelessness is greater than 10 and includes veterans who are currently unhoused.

    The copied features on the map

    Next, you'll create charts to explore the unique characteristics of the homeless population across different communities.

  12. Open the Symbology pane for the Homeless_Tapestry layer. Set the following parameters:
    • For Primary symbology, choose Graduated Colors.
    • For Field, choose Total Homeless People.
    • For Method, choose Natural Breaks (Jenks).
    • For Classes, choose 5.
    • For Color scheme, choose the Red-Purple (Continuous) color scheme.

    Suggested symbology

  13. In the Contents pane, right-click the Homeless_Tapestry layer, point to Create Chart, and choose Scatter Plot.
  14. If necessary, dock the empty chart window (Homeless_Tapestry - Chart of Homeless_Tapestry) below the map.

    Docking the chart below the map

  15. In the Chart Properties pane, on the Data tab, set the X-axis Number parameter to Homeless Veterans and the Y-axis Number to Chronically Homeless People.
  16. Under Statistics, uncheck Show linear trend.

    Chart Properties scatterplot parameters

    The trend line is most interesting when you want to look at correlations. In this case, you're using the scatterplot to explore variable values.

  17. Click the General tab and set the following parameters:
    • For Chart title, type Veterans experiencing homelessness and People experiencing chronic homelessness.
    • For X axis title, type Veterans experiencing homelessness.
    • For Y axis title, type People experiencing chronic homelessness.

    The chart updates automatically. Most of the data points are clustered in the lower left of the chart, with a few outliers in the upper right. This skew is common for many types of data. To improve the visualization of skewed data, you'll use a log axis.

    Scatterplot with an outlier
    Note:

    On a logarithmic scale, increments increase by magnitudes. Logarithmic scales are useful when visualizing data with large positive skew, where the bulk of data points have a small value, and a few data points have very large values. Changing the scale of the axis doesn't change the value of the data, just the way the data values are displayed.

  18. In the Chart Properties pane, click the Axes tab. Under both the X-axis and the Y-axis sections, check Log axis.

    Log axes checked

    Now that you can better see the data points in your chart, you can begin to explore the characteristics of individual tracts through interactive selection.

  19. On the chart, click the census tract with the highest number of veterans experiencing homelessness and the highest number of people experiencing chronic homelessness (the point in the upper right corner).

    A selection is also made on the map, identifying the location of the census tract of interest. Additionally, the symbol color on the map, reflecting the total number of people experiencing homelessness (the Total Homeless People field), matches the symbol color on the chart. Your map and chart are providing information about three different variables, plus location and value magnitude.

    Linked map and scatterplot

    Ending homelessness among veterans is an important priority for Los Angeles. Moving individuals experiencing chronic homelessness into permanent housing is an effective way to decrease overall homeless management costs. By providing aid and support to the people experiencing homelessness in the tract you identified, Los Angeles can address both of these objectives.

Explore homeless characteristics in more detail

To explore the attributes of individual census tracts in more detail, you'll create additional charts that only show data for selected features. You'll begin by exploring different types of shelter options.

  1. In the Contents pane, right-click the Homeless_Tapestry layer, point to Create Chart, and choose Bar Chart.
  2. Dock the bar chart window next to the map so both charts and the map are visible.

    Docked charts around a map

  3. In the empty chart window, for Filter, click Filter By Selection.

    Option to only show selected features in a chart

    Note:

    When Filter By Selection is enabled, the button will appear as blue.

    Now, only selected features appear in the chart.

  4. In the Chart Properties pane, on the Data tab, for Category or Date, choose Community.
  5. For Aggregation, choose Mean.
  6. For the Numeric Field(s) parameter, click Select and choose the following fields that represent shelter types:
    • % Cars, Vans, Campers
    • % Tents, Encampments
    • % Street Homeless
    • % Emergency Shelters, Transitional Housing, Safe Havens

    Chart properties

  7. Click Apply.
  8. Click the Series tab. For the % Emergency Shelters, Transitional Housing, Safe Havens field, change the Label to % Shelters.

    New label

  9. On the General tab, change Chart title to Shelter Type and Y axis title to % People experiencing homelessness .
  10. Select a few census tracts in different locations on the map, in the scatterplot, or in the table.
    Note:

    To select census tracts on the map, click Select on the ribbon in the Selection group. Use the Shift key to add tracts to ones already selected.

    The bar chart updates to show the shelter type breakdown sorted by community.

    Note:

    Your chart may look different than the chart in the example.

    Example bar chart exploring shelter options by community

    This chart allows you to investigate where people experiencing homelessness are finding shelter. This is important. For locations where many people are living in their vehicles, for example, implementing programs like SafeParkingLA may be effective. Preventing trash accumulation in locations with large numbers of encampments is important for preventing disease outbreaks.

    Next, you'll create another bar chart to explore different risk factors contributing to people experiencing homelessness.

  11. In the Contents pane, right-click the Homeless_Tapestry layer, point to Create Chart, and choose Bar Chart.
  12. Dock the new chart window to the left of the scatterplot.
  13. In the empty chart, click Filter By Selection.
  14. In the Chart Properties pane, on the Data tab, change the following parameters:
    • For Category or Date, choose Community.
    • For Aggregation, choose Mean.
  15. For Numeric Field(s), choose the following fields:
    • % Addicted
    • % w Brain Injuries
    • % w Serious Mental Illness
    • % w HIV/AIDS
    • % w Developmental Disabilities
    • % w Physical Disabilities
    • % Victims of Domestic Violence
  16. Click Apply.
  17. On the General tab, change Chart title to Risk Factors and Y axis title to % People experiencing homelessness.
  18. Select a few census tracts on the map, in the scatterplot, or in the table.

    Exploring homeless risk factors and shelter types

    By understanding the primary factors contributing to people becoming homeless tract by tract, Los Angeles can provide targeted homelessness prevention.

Create a histogram

Another way to explore the numeric fields in your data is to create a histogram. Histograms show how data values are distributed. You'll create a histogram to show the distribution of children experiencing homelessness.

  1. In the Contents pane, right-click the Homeless Tapestry layer, point to Create Chart, and choose Histogram.
  2. Dock the newly opened chart window on top of the scatterplot.
  3. In the Chart Properties pane, on the Data tab, set the following parameters:
    • For Number, choose % Under age 18.
    • For With transformation, choose Square Root.
    • For Number of bins, type 40.
    • Uncheck Mean.
    Note:

    The square root transformation improves visualization of skewed data.

    Properties for the histogram

  4. On the Axes tab, under X-axis, for Number format, click the Determine display formatting for numeric field types button.

    Determine display formatting for numeric field types button

  5. Change Category to Numeric, change Decimal places to 2, and click Apply.

    The y-axis (Count) is the number of census tracts associated with different percentages of homeless people under age 18.

  6. On the General tab, change Chart title to Children experiencing homelessness. Change X axis title to % Under Age 18 (sq root).
  7. Select the right tail of the histogram to highlight the census tracts with the highest percentage of people experiencing homelessness under the age of 18.

    Exploring the distribution of homeless children

    Tip:

    You can select individual bars in a bar chart using the Ctrl key or by dragging a box over consecutive bars.

    You can explore other characteristics of the homeless population by changing the histogram variable. You'll do that next.

  8. On the Data tab for the histogram, change Number to % Female.
  9. Select the right tail of the histogram again and examine the location, shelter type, and risk factors for census tracts with the highest percentage of female homeless.

    Your charts and maps identify locations where services for women and children, as well as runaway youth programs, are needed most.

    Understanding differences among people experiencing homelessness from one location to another allows you to be more effective in meeting their needs and providing support. The analyses above focused on census tracts; next, you'll assess homeless characteristics and service needs by community.

Explore homeless characteristics by community

To see results for all the tracts in individual communities, you'll create a bar chart to aggregate community census tracts.

  1. In the Contents pane, right-click the Homeless Tapestry layer, point to Create Chart, and choose Bar Chart.
  2. Dock the new chart over the histogram so all three bar charts and the map are visible.
  3. In the Chart Properties pane, on the Data tab, set the following parameters.
    • For Category or Date, choose Community.
    • For Aggregation, choose Sum.
    • For Numeric Field(s), choose Total Homeless People.
  4. On the General tab, change Chart title to Total people experiencing homelessness by Community and Y axis title to Total people experiencing homelessness.
  5. On the Data tab, under Sort, choose Y-axis Descending.

    Y-axis Descending option under Sort in the Chart Properties pane

  6. Select the two communities with the largest number of homeless people - Skid Row and Hollywood. Compare the associated shelter types and risk factors.
    Note:

    Point to a bar in the histogram to see the community name.

    Exploring the characteristics of homeless communities

    From the charts you have created, you see that the Skid Row community has more than twice as many people experiencing homelessness as the Hollywood community. The Shelter Type chart reveals that very few people who are unhoused in Skid Row live in cars, vans, or campers. The Risk Factors chart indicates that a large percentage of people experiencing homelessness in Hollywood are victims of domestic violence.

  7. Close the charts, clear any active selections, and save the project.

You worked with point-in-time count data for the City of Los Angeles and Los Angeles County. Using maps and charts, you identified locations where people experiencing homelessness concentrate and explored variations in homeless community characteristics. You can use this same workflow with your own homeless census data.

If you're looking for resources to help you improve your survey, or if you need technology to help you collect the census data, consider the following suggestions:

  • A key to prevention is understanding where people are becoming homeless. The point-in-time counts tell you where people experiencing homelessness are but also provide the opportunity to ask about last permanent residence (ZIP Code, address, or street intersection) and perceived reasons for homelessness to learn more about risk factors.
  • Poor transitions from foster care, jail, and other institutions may be an important factor contributing to a person becoming homeless. You can record if a person's last permanent housing was an institution and use this information to design appropriate remediation.
  • Los Angeles County makes its survey questions public. Looking at its survey may help you design or enhance your own survey questions.
  • A number of resources are available to help you learn about using Esri technology to conduct your own point-in-time homeless census. The Survey123 for ArcGIS - Homeless Count video, in particular, explains the point-in-time count application.
  • Los Angeles doesn't make the raw survey results public; only regional estimates are provided for many variables (age, risk factors, race/ethnicity, gender, veteran status, and so on). As a consequence, apportionment was applied to get tract-level estimates. You may notice many of the estimates are duplicated across tracts. This is because, apparently, the apportionment algorithm aligns with the original estimation method used. If you have raw survey results, however, you won't need to apportion the survey data but can instead use the survey results directly. The workflow outlined above will work fine.

Next, you'll consider where to invest in new resources to support and aid people experiencing homelessness. This is often a contentious issue. The workflow that follows encourages transparency and uses spatial overlay analysis to promote consensus.


Promote consensus

Previously, you identified where people experiencing homelessness concentrate and explored the characteristics of different homeless communities. Next, you'll answer the question, Where are the best locations to invest in new resources to aid and support people experiencing homelessness?

Ask ten stakeholders and you'll likely get ten different ideas about where to invest in new resources. There is no one best solution. Mapping planning scenario objectives, however, is an effective way to provide neutral ground for discussion, to encourage both transparency and collaboration, and to increase focused engagement.

You'll create several potential planning scenarios, each promoting a different objective. You'll then overlay all of the planning scenario maps to see if any locations optimize more than one objective.

Add resources to optimize social equity

First, you'll create a map that shows where to add resources if the goal is to optimize social equity. You'll compute a supply (available shelters) versus demand (proportional of population experiencing homelessness) variable relating the total county population in each tract to the total population experiencing homelessness in each tract. This value will show which census tracts currently have a proportional amount of resources for the proportion of people experiencing homelessness in that tract.

  1. If necessary, open the Homelessness project in ArcGIS Pro.
  2. In the Contents pane, turn off all layers except the County Outline layer, Additional Resources layer, and basemap.
  3. Drag the County Outline and Additional Resources layers to the top of the list of layers.

    To compute the social equity index, you'll need the total number of residents and the total number of people experiencing homelessness countywide. Later in the lesson, you'll need the total number of sheltered homeless countywide, so you'll also calculate that.

  4. In the Geoprocessing pane, search for and open the Summary Statistics tool.
    Note:

    Learn more about the Summary Statistics tool.

  5. In the Summary Statistics tool pane, set the following parameters:
    • For Input Table, choose Additional Resources.
    • For Output Table, type Sum_Pop_Variables.
    • For Field, choose 2017 Total Population, Total Homeless People, and Total Sheltered People and ensure Statistic Type is set to Sum for all three fields.

    Summary Statistics parameters

  6. Click Run.

    The tool runs. It writes the sum values for each field to the output table.

  7. In the Contents pane, right-click Sum_Pop_Variables and open the table.
    Note:

    It is in the Standalone Tables section of the Contents pane.

    The total residential population is 9,446,500. The total population of people experiencing homelessness is 49,703 (rounded up from 49,702.811; the floating-point number indicates some counts were estimated). The total number of people experiencing homelessness who are temporarily sheltered is 12,933.

  8. Close the table.

    Next, you'll add a field and calculate the social equity index.

  9. In the Geoprocessing pane, open and run the Add Field tool with the following parameters:
    Note:

    Learn more about the Add Field tool.

  10. In the Add Field tool pane, set and run the following parameters:
    • For Input Table, choose Additional Resources.
    • For Field Name, type SocialEquityIndex.
    • For Field Type, choose Float (32-bit floating point).
    • For Field Alias, type Social Equity Index.

    Add Field tool parameters

    Now that you have added the field, you will calculate it.

  11. In the Geoprocessing pane, search for and open the Calculate Field tool.
    Note:

    Learn more about the Calculate Field tool.

  12. In the Calculate Field tool pane, set and run the following parameters:
    • For Input Table, choose Additional Resources.
    • For Field Name, choose Social Equity Index.
    • Under Expression, for SocialEquityIndex =, type or copy and paste (!TOTPOP_CY! / 9446500) - (!totPeople! / 49703).

    Calculate Field parameters set.

    Note:

    Tool expressions often use field names enclosed in exclamation marks rather than field aliases. The field name for the 2017 Total Population field is TOTPOP_CY; the field name for the Total Homeless People field is totPeople.

    The first ratio in the expression (!TOTPOP_CY! / 9446500) reflects the portion of the county's residents in each census tract. The second ratio (!totPeople! / 49703) is the proportion of people experiencing homelessness in each census tract.

    These two ratios are subtracted. Complete equity would mean the difference of the ratios is zero and every census tract in the county has the same proportion of people experiencing homelessness for the proportion of total residents in that tract.

    If the difference between the two ratios is a positive value (a tract with 4 percent of the county's residents and 2 percent of the county's population of people experiencing homelessness), it shows which census tracts have a proportional amount of resources for the people experiencing homelessness in that tract. Adding new resources in these areas will help spread the shared responsibility of providing resources to the homeless population more equitably across the whole county.

    If the difference between the two ratios is a negative value, it means the proportion of people experiencing homelessness in that census tract is higher relative to its proportion of total residents within the tract. These tracts are likely already carrying more responsibility to provide services to people experiencing homelessness.

  13. Open the Symbology pane for the Additional Resources layer. In the Symbology pane, set the following parameters:
    • For Primary symbology, choose Graduated Colors.
    • For Field, choose Social Equity Index.
    • For Method, choose Standard Deviation.
    • On the Classes tab, click More and choose Reverse values.
    Note:

    If you don't see the Social Equity Index field as an option for the Field parameter, change Primary symbology to Unique Values. Then change it back to Graduated Colors to refresh the field list.

    Reversing the symbol order and values puts the class you want to focus on at the top of the legend. The positive deviations from the mean reflect tracts where new resources should be added.

    Classes tab in the Symbology pane

  14. For the 0.50 - 0.84 Std. Dev. class break, right-click the symbol and choose Peony Pink (to see the name of a color, point to the color). For the -0.50 - 0.50 Std. Dev class break, right-click the symbol and choose Arctic White.

    Always use a neutral color for the -0.50 - 0.50 Std. Dev. class break because it represents the mean (what's expected and, consequently, not most interesting).

  15. For the remaining class breaks, choose progressively darker gray colors (Gray 10%, Gray 30%, Gray 50%).

    Symbol colors updated in the Classes tab

  16. Click More and choose Format all symbols.
  17. Click the Properties tab and change Outline color to Gray 10%. Change Outline width to 0.25 and click Apply.

    Next, you will update your symbology labels to reflect a tract's current and future opportunity for sharing responsibility to support the county's homeless community. Areas with a lower proportion of the homeless population relative to their housed population could add resources for the homeless community to contribute toward the county's shared responsibility.

  18. Click the back button, and in the Classes table, change the labels to the following values:
    • For 0.50 - 0.84 Std. Dev, type Add new resources to promote equitable support of the homeless community.
    • For -0.50 - 0.50 Std. Dev, type Supporting average proportion of the County's homeless community.
    • For -1.50 - -0.50 Std. Dev, type Supporting more of the County's homeless community.
    • For -2.50 - -1.50 Std. Dev, type Supporting much more of the County’s homeless community.
    • For < -2.50 Std. Dev, type Supporting most of the County’s homeless community.

    Updated labels

    The symbology updates on the map and so does the legend in the Contents pane.

    Optimize social equity

    The tracts where new resources will promote social equity have index values larger than 0.000862. To flag these tracts, you'll create a field that has two possible values, 0 and 1. You'll give tracts with an index less than 0.000862 a value of 0 and tracts with a higher index a value of 1. You'll sum the flag fields at the end of this lesson to identify locations that optimize multiple planning scenario objectives.

  19. In the Geoprocessing pane, search and open the Add Field tool.
    Note:

    The tool will also appear on the Favorites tab in the Recent list so you don't have to search for it every time.

  20. Run the tool with the following parameters:
    • For Input Table, choose Additional Resources.
    • For Field Name, type SocialEquityFlag.
    • For Field Type, choose Short (16-bit integer).
  21. In the Geoprocessing pane, search and open the Calculate Field tool. Run the tool with the following parameters:
    • For Input Table, choose Additional Resources.
    • For Field Name, choose SocialEquityFlag.
    • Under Expression, for SocialEquityFlag = type 0.

    SocialEquityFlag = 0 expression

  22. In the Geoprocessing pane, search and open the Select Layer By Attribute tool.

    Note:

    Learn more about the Select Layer By Attribute tool.

  23. Run the tool with the following parameters:
    • For Input Rows, choose Additional Resources.
    • For Selection type, choose New selection.
    • For Expression, create the expression Where Social Equity Index is greater than 0.000862.

    Selecting tracts with the largest social equity indices

    Four tracts are selected.

  24. Open and run the Calculate Field tool with the following parameters:
    • For Input Table, choose Additional Resources.
    • For Field Name, choose SocialEquityFlag.
    • Under Expression, for SocialEquityFlag = type 1.

    For the selected records, the value of the SocialEquityFlag field is set to 1.

    You now have a field that identifies the tracts in the county with a lower than equal proportion of people experiencing homelessness. With this field, you can more easily flag census tracts that have opportunity to increase their shared responsibility of supporting the county's homeless community.

  25. Clear the selection.

    You'll save your symbology for the Social Equity Index field to a file you can view later. Later, you'll override an existing symbology with this saved symbology.

  26. In the Contents pane, right-click the Additional Resources layer, point to Sharing and click Save as Layer File.
  27. In the Save as Layer File window, specify a folder and name the file SocialEquityMap.lyrx. Click Save.
  28. Save the project.

Add resources to prioritize accessibility

A second possible planning scenario for locating new resources is to add them so they prioritize accessibility for existing homeless people. For this option, you'll compute another supply versus demand variable, but you'll base it on the relationship between the existing homeless population and existing homeless resources. Because Los Angeles doesn't yet provide complete data about homeless resources, you'll use the number of sheltered homeless people as a surrogate for all homeless resources.

You'll add a new field, calculate the accessibility index, and create a map.

  1. Open and run the Add Field tool to add a field to the Additional Resources layer named AccessibilityIndex of type Float (32-bit floating point). For Field Alias, type Accessibility Index.
    Note:

    If you have trouble adding a field, review previous steps using the Add Field tool.

  2. Use the Calculate Field tool and calculate the values for the Accessibility Index field using the following parameters:
    • For Input Table, choose Additional Resources.
    • For Field Name, choose Accessibility Index.
    • For Expression, type (!totPeople! / 49703) - (!totSheltPeople! / 12933).
    Note:

    If you have trouble calculating a field, review the step to calculate a field in the previous section.

    The first ratio (!totPeople! / 49703) is the proportion of all people experiencing homelessness in a tract. The second ratio is the proportion of all sheltered people experiencing homelessness in a tract.

    Equitable accessibility is exhibited in tracts where the proportion of homeless people matches the proportion of sheltered homeless people. In Los Angeles, there is a large deficit of shelter opportunities and the formula takes this into account. A tract with 2 percent of the county's homeless people should have 2 percent of the county's available shelter opportunities. That is, the deficit should be justly distributed across the county.

    If the ratios are subtracted and the difference is zero, this difference represents an optimal distribution of existing shelters. If the difference is positive, the proportion of homeless people is larger than the proportion of sheltered homeless people, indicating inadequate access to shelter opportunities.

    If the difference is negative, the proportion of shelter opportunities is larger than the proportion of homeless people, indicating better than average access to shelter opportunities.

    You'll symbolize the Accessibility Index field the same way you symbolized the Social Equity Index field.

  3. In the Contents pane, right-click Additional Resources and click Symbology. In the Symbology pane, for Field, choose Accessibility Index.
  4. On the Classes tab, right-click the symbol colors for the class breaks with positive deviations from the mean and choose the following colors:
    • For > 2.5 Std. Dev., choose Peony Pink.
    • For 1.50 - 2.5 Std. Dev., choose Fuchsia Pink.
    • For 0.50 - 1.50 Std. Dev., choose Rhodolite Rose.

    Positive standard deviation classes set to shades of pink

  5. For -0.50 - 0.50 Std. Dev., choose a neutral color, like Arctic White.
    • For -1.5 - -0.5 Std. Dev, choose Gray 10%.
    • For -2.5 - -1.5 Std. Dev, choose Gray 30%.
    • For < -2.5 Std. Dev, choose Gray 50%.
  6. In the Label column, type the following:
    • For the first symbol, type Add new resources here to improve accessibility.
    • For the fourth symbol, type Balanced.
    • For the seventh symbol, type Best access to existing resources.
    • Remove all other labels.

    Label column updated in the Classes tab

  7. Click More and choose Format all symbols.
  8. Change Outline color to Gray 10% and Outline width to 0.25. Click Apply.

    The symbology updates on the map.

    Prioritize accessibility

    The tracts with accessibility indices greater than 0.001315 are potential candidates for new resources if the goal is to improve accessibility in an equitable manner. You'll create a field to flag these tracts.

  9. Use what you have learned to use the Add Field tool and add a field to the Additional Resources layer named AccessibilityFlag with Field Type set to Short (16-bit-integer).
  10. Use the Calculate Field tool so all the values for the AccessibilityFlag field in the Additional Resources table are 0
  11. Open and run the Select Layer By Attribute tool to select tracts from the Additional Resources layer with an Accessibility Index greater than 0.001315.
  12. For the selected records, use the Calculate Field tool to set the AccessibilityFlag values to 1.
  13. Clear the selection.
  14. In the Contents pane, right-click the Additional Resources layer, point to Sharing, and choose Save as Layer File. Browse to a folder of your choice and save the file with the name AccessibilityMap.lyrx.
  15. Save the project.

Add resources to prioritize highest risk areas

A third possible planning scenario is to focus on areas at highest risk for people becoming homeless. If adequate resources are made available in such areas, people are more likely to remain housed and continue living in their own communities where there is a stronger likelihood of access to personal and social resources. Providing new resources in high-risk areas is the strategy being promoted by the mayor of New York.

You'll use the Risk_Surface layer you created earlier in the lesson to indicate how likely a tract is to have people at risk for becoming unhoused.

  1. In the Geoprocessing pane, open and run the Add Field tool using the following parameters:
    • For Input Table, choose Additional Resources.
    • For Field Name, type RiskIdx.
    • For Field Type, choose Float (32-bit floating point).
    • For Field Alias, type Risk Index.

    You'll link the records in the Risk_Surface layer with records in the Additional Resources layer. First, you'll remove any attribute indexes from the Risk_Surface layer to ensure there are no issues with the join.

  2. Open and run the Add Join tool with the following parameters:
    • For Input Table, choose Additional Resources.
    • For Input Join Field, choose TRACT.
    • For Join Table, choose Risk_Surface.
    • For Join Table Field, choose TRACT.
    • Confirm that Keep All Target Features is checked.

    Add Join tool parameters

    Note:

    Learn more about the Add Join tool.

    You'll transfer the risk rankings from the Risk_Surface layer to the Additional Resources layer. The Render Rank (LABELRANK) field has rankings that range from 1 (highest risk) to 2,343 (lowest risk). To be consistent with the other indices you've been using, you'll reverse the rankings so that the largest rank reflects the highest risk.

  3. Open and run the Calculate Field tool with the following parameters:
    • For Input Table, choose Additional Resources.
    • For Field Name, choose Additional_Resources.RiskIdx.
    • For Expression, type 2344 - !Risk_Surface.LABELRANK!.

    Calculating the risk index

    Each field identifies its source layer. Next, you'll unlink the joined layers.

  4. In the Contents pane, right-click Additional Resources, point to Joins and Relates, and choose Remove All Joins. Click Yes to confirm.
    Note:

    Learn more about Remove Joins.

    Next, you'll symbolize the layer by the Risk Index field. Since risk isn't a diverging attribute, it's appropriate to use graduated colors.

  5. Open the Symbology pane for the Additional Resources layer and set the following parameters:
    • For Field, choose Risk Index.
    • On the Classes tab, right-click the first symbol and choose Cattleya Orchid, the second symbol to Fuchsia Pink, the third symbol to Rhodolite Rose, and the fourth and fifth symbols to Arctic White.
    • In the Label column, change the first label to Highest risk for becoming homeless, the third label to Medium risk for becoming homeless, and the fifth label to Lowest risk for becoming homeless. Remove the other labels.

    Standard deviation rendering, risk rankings

  6. Click More and choose Format all symbols. On the Properties tab, for Outline color, choose 10% Gray and change Outline width to 0.25. Click Apply.

    The symbology updates on the map.

    Focus on the highest-risk areas

    The tracts with risk indices greater than 1,513 are good candidates for investment if the goal is to provide new resources where people are at highest risk for becoming unhoused.

  7. Use what you have learned to create and calculate a field named RiskFlag to the Additional Resources layer.
    • Open and run the Add Field tool, set Input Table to Additional Resources, Field Name to RiskFlag, and Field Type set to Short (16-bit integer).
    • Run the Calculate Field tool to set all values for RiskFlag to 0.
    • Run the Select Layer by Attribute tool to select tracts with Risk Index is greater than 1513.
    • Run the Calculate Field tool to set all the selected tracts for RiskFlag to a value of 1.
  8. Clear the selection.
  9. In the Contents pane, right-click the Additional Resources layer, point to Sharing, and choose Save as Layer File. Save the file with the name RiskMap.lyrx.
  10. Save the project.

Centralize homeless resources

A fourth possible planning scenario is to consolidate new resources into triage centers by encouraging centralization. This creates resource hubs and is the model San Francisco has adopted. For this planning scenario, you'll use the total number of sheltered people experiencing homelessness as a surrogate for existing homeless resources.

  1. Open the Symbology pane for the Additional Resources layer. Symbolize the layer using the following parameters:
    • For Primary symbology, choose Graduated Colors.
    • For Field, choose Total Sheltered People.
    • For Method, choose Standard Deviation.
    • Change the first symbol to Cattleya Orchid, the second symbol to Fuchsia Pink, the third symbol to Rhodolite Rose, and the fourth symbol to Arctic White.
    • For the Label column, change the first row to Highest priority, the second label to Medium priority, the third label to Low priority, and the fourth label to Lowest priority.

    Symbology pane for Additional Resources showing Total Sheltered People and symbology classes configured

  2. Use what you have learned to format all symbols to have a 0.25 point light gray outline.

    The map updates to show the new symbology.

    Consolidate new resources

    The tracts with resource counts larger than 85 are good candidates for new resources if the goal is to promote centralization of resources into navigation hubs.

  3. Use what you have learned to create and calculate a CentralizationFlag field to the Additional Resources layer.
    • Open and run the Add Field tool, set Input Table to Additional Resources, Field Name to CentralizationFlag, and Field Type to Short (16-bit integer).
    • Run the Calculate Field tool to set all values for CentralizationFlag to 0.
    • Run the Select Layer by Attribute tool to select tracts with Total Sheltered People is greater than 85.
    • Run the Calculate Field tool to set all the selected tracts for CentralizationFlag to a value of 1.
  4. Clear the selection.
  5. In the Contents pane, right-click the Additional Resources layer, point to Sharing, and choose Save as Layer File. Save the file with the name CentralizationMap.lyrx.
  6. Save the project.

Focus on the most vulnerable homeless populations

Another option is a street strategy that identifies the most vulnerable homeless populations, evidenced by 311 calls, crime incidents involving people experiencing homelessness, and a large number of individuals experiencing chronic homelessness. These locations become candidates for rapid-response, focused interventions aimed at providing every homeless person with precisely what they need to move out of homelessness permanently. Some research papers conclude that a small portion of the homeless population (primarily the chronically homeless street population) use a majority of the money targeted for homelessness.

Note:

311 is a phone hotline number that United States residents can use to receive local resource information or to report a problem in their community. This analysis uses the number of calls relating to people experiencing homelessness or encampments where homeless people are staying.

The data needed for this analysis isn't available for all tracts in Los Angeles County. You'll create the planning scenario map for tracts where data is available within the City of Los Angeles. Similar to the workflow for creating a risk surface layer, you'll create a street strategy target tract with worst-case values and use the Similarity Search tool to rank all other tracts in relation to that target. Finally, you'll symbolize the result layer and create a flag variable, like you've done for each of the other planning scenarios.

  1. Turn off the County Outline layer and use the Select tool to select any tract to serve as the street strategy target tract.

    A selected census tract

  2. Open and run the Copy Features tool using the following parameters:
    • For Input Features, choose Additional Resources.
    • For Output Feature Class, type SS_Target_Tract.

    The copied layer, showing only the selected tract, is added to the map.

  3. Turn off the Additional Resources layer.
  4. Use the Select tool to select the copied census tract.
  5. On the ribbon, click the Edit tab. In the Tools group, click Move.
  6. Drag the selected census tract outside of Los Angeles County.
  7. At the bottom of the map, click Finish.
  8. On the ribbon, in the Manage Edits group, click Save. In the Save Edits window, click Yes.
  9. Clear the selection. Turn on the County Outline and Additional Resources layers.

    Next, you will fill out the SS_Target_Tract table with the highest risk values for all census tracts in the county.

    SS_Target_Tract created and moved outside the county border as a reference

  10. Open the SS_Target_Tract attribute table and edit the field values as follows:
    • For VictimCount, type 332.
    • For 311 Call Count, type 330.
    • For Chronically Homeless, type 854.
    Tip:

    To find the largest values, open the attribute table for the Additional Resources layer and sort the field columns to view the highest risk value for each field.

    Highest risk values set for the SS_Target_Tract table

  11. On the ribbon, on the Edit tab, in the Manage Edits group, click Save. In the Save Edits window, click Yes.

    You'll rank all tracks that have these data fields, from worst to best.

  12. Search for and open the Similarity Search tool and set the following parameters:
    • For Input Features To Match, choose SS_Target_Tract.
    • For Candidate Features, choose Additional Resources.
    • For Output Features, type Street_Strategy_Rankings.
    • Ensure that Collapse Output To Points is unchecked.
    • For Number Of Results, type 0.
    • For Attributes Of Interest, choose VictimCount, 311 Call Count, and Chronically Homeless.

    Similarity Search tool parameters set

  13. Click Run.
    Note:

    You may receive a warning about problems reading some records. This is expected because the attribute values are null for tracts where data isn't available. You can ignore the warnings.

  14. In the Contents pane, turn off the Additional Resources and SS_Target_Tract layers.

    The resulting map shows census tracts in the City of Los Angeles, one of the many jurisdictions within Los Angeles County, symbolized by how similar the tract is to SS_Target_Tract with the highest vulnerability values for three vulnerability risk factors.

    Resulting map ranking census tracts in the City of Los Angeles with the highest risk factors for the street strategy

    Next, you'll add a Render Rank field in the output layer to create a street strategy index field. First, you'll create a field for the street strategy index values.

  15. Open and run the Add Field tool using the following parameters:
    • For Input Table, choose Additional Resources.
    • For Field Name, type SSIndex.
    • For Field Type, choose Float (32-bit floating point).
    • For Field Alias, type Street Strategy Index.

    You'll join the output from the Similarity Search tool to the Additional Resources layer.

  16. Open and run the Add Join tool using the following parameters:
    • For Input Table, choose Additional Resources.
    • For Input Join Field, choose OBJECTID.
    • For Join Table, choose Street_Strategy_Rankings.
    • For Output Join Field, choose CAND_ID.
    • Uncheck Keep All Target Features.

    Add Join tool parameters set

    There are 940 features with rankings from 1 (most similar to the most vulnerable value) to 940 (least similar to the most vulnerable value). You'll create an index field reversing these rankings, so the largest number reflects the highest-priority intervention site.

  17. Open and run the Calculate Field tool using the following parameters:
    • For Input Table, choose Additional Resources.
    • For Field Name, choose Additional_Resources.SSIndex.
    • Under Expression, for Additional_Resources.SSIndex, type 941 - !Street_Strategy_Rankings.LABELRANK!.

    Calculate Field tool to calculate the SSIndex field

  18. Open and run the Remove Join tool to remove the join from the Additional Resources layer.
  19. In the Contents pane, turn off all layers except the County Outline, Additional Resources, and basemap layers.
  20. For the Additional Resources layer, open the Symbology pane and symbolize the Street Strategy Index field in a similar way to how you symbolized the other index fields.
    • For Field, choose Street Strategy Index.
    • In the Classes table, right-click the first symbol and choose Cattleya Orchid. For the second symbol, choose Fushia Pink. For the third symbol, choose Rhodolite Rose.
    • For the fourth and fifth symbols, choose Arctic White.
    • Click More and Format all symbols. Change the Outline width to 0.25 and the Outline color to Gray 10%.
    • Click Apply.

    The symbology updates on the map.

    Priority intervention sites

  21. In the Symbology pane, click the back button, and on the Classes tab, update the Label column to the following:
    • Highest priority for intervention site
    • High priority
    • Medium priority
    • Low priority
    • Lowest priority

    Based on the legend, the tracts with a Street Strategy Index value larger than 606 are the best candidates for focused intervention.

    Street strategy legend

  22. Use what you have learned to add and calculate a StreetStrategyFlag field in the Additional Resources layer.
    • Open and run the Add Field tool, set Input Table to Additional Resources, Field Name to StreetStrategyFlag, and Field Type to Short (16-bit integer).
    • Run the Calculate Field tool to set all values for StreetStrategyFlag to 0.
    • Run the Select Layer by Attribute tool to select tracts with Street Strategy Index is greater than 606.
    • Run the Calculate Field tool to set all the selected tracts for StreetStrategyFlag to a value of 1.
  23. Clear the selection. Save the symbology as a new file named StreetStrategyMap.lyrx.

Identify locations that optimize multiple objectives

Next, you'll perform a spatial overlay to identify tracts providing opportunities to address more than one planning scenario objective.

You set the flag fields to 1 when a planning scenario objective was met. You'll create a field to sum the number of objectives met for each tract.

  1. Open and run the Add Field tool with the following parameters:
    • For Input Table, choose Additional Resources.
    • For Field Name, type NumObjsMet.
    • For Field Type, choose Short (16-bit integer).
    • For Field Alias, type Number of Objectives Met.

    Next, you'll calculate the sum.

  2. Open and run the Calculate Field tool with the following parameters:
    • For Input Table, choose Additional Resources.
    • For Field Name, choose Number of Objectives Met.
    • Under Expression, for NumObjMet =, type !SocialEquityFlag! + !AccessibilityFlag! + !RiskFlag! + !CentralizationFlag! + !StreetStrategyFlag!.

    Calculate Field tool parameters for calculating the Number of Objectives Met field

  3. Open the Symbology pane for the Additional Resources layer and set the following parameters:
    • For Primary symbology, choose Graduated Colors.
    • For Field, choose Number of Objectives Met.
    • For Method, choose Manual Interval.
    • If necessary, on the Classes tab, change the first symbol to Cattleya Orchid, the second symbol to Fuchsia Pink, the third symbol to Rhodolite Rose, and the fourth and fifth symbols to Arctic White.
    • If necessary, format all symbols to have a 0.25 point light gray outline.

    Symbology pane parameters for the Number of Objectives Met field

  4. On the Classes tab, in the Upper value column, update the following:
    • For the first symbol, type 4.0.
    • For the second symbol, type 3.0.
    • For the third symbol, type 2.0.
    • For the fourth symbol, type 1.0.
    • For the fifth symbol, type 0.8.
  5. In the Label column, change the first label to 4, the second label to 3, the third label to 2, the fourth label to 1, and the fifth label to 0.

    The break values and labels are updated.

    Symbology for map showing the number of objectives met in each tract

    The darkest areas on the map meet the largest number of planning scenario objectives.

    Optimizing multiple objectives

In this lesson, you created planning scenario maps optimizing different objectives. You then used spatial overlay to identify locations meeting multiple planning scenario objectives.

Only a few of the many possible planning scenarios are presented in this lesson. Mapping each one is important, not only for the spatial overlay at the end of the workflow, but also because maps provide an excellent foundation for open communication, collaboration, and building consensus.

The street strategy planning scenario presented here is powerful because it prioritizes intervention for the highest-crime, highest-complaint, and highest-cost locations. For this lesson, the analysis used 311 complaint data relating to homeless people or encampments, crime data relating to victims who are experiencing homelessness, and point-in-time count data for the number of people experiencing chronic homelessness. If you have access to emergency medical system (EMS) call data for individuals experiencing homelessness or crime incidents involving homeless perpetrators, the map for this planning scenario will more accurately reflect the most vulnerable homeless populations and locations where addressing homelessness can have the biggest impact on costs.

This lesson provides a glimpse into the homelessness crisis taking place throughout Los Angeles County. It outlines strategies to help cities, agencies, and spatial analysts understand, prevent, and manage homelessness. Key to prevention is knowing where people are at risk of becoming homeless.

By completing each workflow, you learned how to analyze risk factors to create a risk map for people becoming homeless. You also learned a strategy for prioritizing locations for prevention programs. Using charts and maps, you explored differences among people experiencing homelessness across the neighborhoods of Los Angeles. Understanding how shelter options and the factors contributing to homelessness vary puts you in a better position to recommend targeted aid and support.

Finally, you created maps showing planning scenarios for resource investment for those experiencing homelessness. Overlaying the planning scenarios revealed which locations that meet multiple planning scenario objectives. These, and other workflows, will be needed to effectively end homelessness.

You can find more lessons in the Learn ArcGIS Lesson Gallery.