Find potential hurricane shelters

During Hurricane Harvey, shelters struggled to harbor the 300,000 people who sought refuge. To prepare for the next hurricane, you want to help the city by identifying more potential hurricane shelters in underserved areas. Using ArcGIS Business Analyst, you'll perform a suitability analysis that weights proximity to existing shelters and demographics data, and accounts for structural and humanitarian guidelines set by the Federal Emergency Management Agency (FEMA).

This tutorial was last tested on November 21, 2022.

View final result

Requirements

• ArcGIS organizational account with a Publisher, Facilitator, or Administrator role in an ArcGIS organization (see options for software access)
• ArcGIS Business Analyst Web App
• 20 credits

Identify potential shelter locations

During Hurricane Harvey, there was not enough space in shelters for everyone escaping the widespread flooding. ArcGIS Business Analyst Web App can be used to determine both potential shelter locations and populations in need. You'll create a project in Business Analyst and add data to it, including existing shelter sites and areas of the city designated as likely to flood. Then, you'll add the locations of schools within the city, as these are facilities that are easy to convert into shelters. Finally, you'll begin narrowing down the schools. Shelters shouldn't be too small, and they should be relatively safe from floodwaters, usually sitting outside of both evacuation zones and FEMA-designated floodplains.

Create a project

Projects are the way Business Analyst organizes and groups together relevant content. For this project, you'll be using point and site data as well as ArcGIS Online layers. These can be added in several ways and are stored in the project pane.

1. Sign in to Business Analyst. If necessary, on the ribbon, click Home.

   The Business Analyst home page opens. On this page, you can see recent updates and a quick-start video, as well as your own content. If you have used Business Analyst before, your latest content will be displayed at the bottom of the screen.

   Note:

   Business Analyst requires a licensed ArcGIS organizational account. If you are the administrator of your organization, you can assign yourself a license. Otherwise, you'll need to contact your administrator for permission.

2. Click Create New Project.

   

   The Create Project window opens.

3. Name the project Emergency Shelter Locations and add your name or initials. Click Create.

   

   A window opens, explaining that a new project is being created. The process might take a few minutes. Once the project is complete, another window opens for confirmation.

4. In the confirmation window, check the box for Open new projects as soon as they are created and click OK.

   

   The list of projects has been updated to show your new project. Depending on your settings, the new project may be opened automatically.

5. If Emergency Shelter Locations is not automatically opened, point to the card and click Open.

   Your new project opens. A pop-up may appear, offering a guided tour of the user interface.

6. If necessary, close the Step-by-Step Guided Tours pop-up.

   

Add data

Next, you'll add data to your new project. There are several ways to add data. The first is a spreadsheet of current shelter locations in Houston. The addresses are saved in a spreadsheet, which you'll upload to Business Analyst. Then, you'll add web maps of evacuation zones and floodplains.

1. Download the houston-shelters XLSX file.
2. On the ribbon, click Add Data and choose Import File.

   

   The Import File pane appears in place of the project pane.

3. For Select a file to import, click Browse. Browse to and double-click the houston-shelters XLSX file.
4. Click Import.

   

   Next, you are prompted to choose the type of data being imported. The options are point locations and geographic boundaries. The hurricane shelters are point locations.

5. For Select the type of data being imported, confirm that Point locations is chosen and click Next.

   

   Next, you'll select the fields in the XLSX file that will determine the location of the shelters on the map.

6. For Confirm the fields match the columns in your file, for Address or Place, choose Street Address. Confirm that City and ZIP are chosen (State can be left empty) and click Next.

   

   The map zooms to Houston, Texas, and points corresponding to hurricane shelters are added to the map. In the Import File pane, you can choose how the points are styled. For now, you'll use the default style.

7. In the Import File pane, for Style the points, click Next.

   Finally, you'll save the points to the project.

8. Confirm that the Save only points option is selected.

   

9. Click Save.
10. After the layer is saved, click I'm Done.

    

    During Hurricane Harvey, there were 38 shelters available. Depending on the map scale you're using, the shelters might be displayed using clustering. Due to proximity, several shelters can be represented by a single pin with the number of points in a box to the upper left. This feature is scale dependent and will disappear as you zoom closer.

    The display shows the Streets basemap, which is good for to better understand the way civilians navigate the city.

11. On the ribbon, click the Emergency Shelters Locations tab.

    

    The project pane opens. The houston-shelters layer has been saved under Other Layers. The check box next to the layer indicates that it is currently being drawn on the map.

    

12. Uncheck the houston-shelters layer.

    The layer is still available for analysis, but it doesn't show on the map. Next, you'll add two web maps with the other data you'll need for your analysis. One shows FEMA-designated floodplains. The other shows evacuation zones for the city. Because these areas are prone to flooding, no shelters should be located within them.

13. On the ribbon, click Add Data and choose Web Maps and Layers.

    The Web Maps and Layers window appears.

14. Click the ArcGIS tab.
15. In the search bar, type USA Flood Hazard Areas and press Enter.

    

    Multiple results include USA Flood Hazard Areas in their names.

16. Point to each result. For the USA Flood Hazard Areas result that is a feature service published by esri, click Open.

    

    The result is added to the map.

    

    This layer is the vector version of United States flood hazard data. The Federal Emergency Management Agency (FEMA) produces Flood Insurance Rate maps and identifies Special Flood Hazard Areas as part of the National Flood Insurance Program's floodplain management. Special Flood Hazard Areas have regulations that include the mandatory purchase of flood insurance. This layer is useful when working on a national scale, but when you're looking at a relatively small area, like Houston, 30 meters can be the difference of an entire building.

    Next, you'll add the evacuation zones layer.

17. Close the Legend window.
18. In the project pane, uncheck USA Flood Hazard Areas to turn it off.
19. On the ribbon, click Add Data and choose Web Maps and Layers.
20. In the search bar, clear the existing text, paste the URL for a web map of evacuation zones in Houston. (https://services2.arcgis.com/j80Jz20at6Bi0thr/arcgis/rest/services/Harris_County_Hurricane_Evacuation_Zones/MapServer), and press Enter.

    The web map is added to the project. If you know the URL or ID of your web map, you can use the Web Maps and Layers window to locate and add it.

    

    The evacuation zone layer shows evacuation data for all of Harris County, where Houston is located.

21. Close the Legend window.
22. On the ribbon, in the search bar, click the arrow and choose Add a boundary.

    Note:

    If you don't see a search bar, resize your window.

    

23. In the search bar, type Houston, TX and press Enter.

    A list of search results appears.

    

    Because you want to analyze data in Houston, the best match is Houston City, Texas.

24. Click Houston City, TX.

    The Houston city boundary is added to the map.

    

25. Close the Houston City, TX pop-up.

    Now that you have the city limits, you'll identify parts of the city that lie within evacuation zones and floodplains. Facilities in flood-prone places can immediately be removed from your list of shelter candidates.

    Note:

    Business Analyst automatically saves your project. If you need to return to it, click the Maps tab on the home page and choose the Emergency Shelter Locations project card. Projects do not save maps, though. To redraw your map, click the Emergency Shelter Locations tab on the ribbon and turn on layers.

Map school locations

The last dataset you need is the locations of all the elementary and secondary schools in Houston. From this list, you'll start identifying potential shelters using visual analysis of schools that lie within areas of likely flooding. Locations should not be considered for shelters if they overlap with hurricane evacuation zones or FEMA-designated floodplains.

1. On the ribbon, click Create Maps and choose Points of Interest Search.

   

   The Points of Interest Search pane appears.

2. Click More options.

   

   Now in addition to being able to search by name, you can search using other options. One option is North American Industry Classification System (NAICS)-based code, a system of facility classification that the United States Census Bureau uses to classify industries. Using this code results in a more accurate search because it narrows the results with much greater accuracy than a keyword search.

3. Click Search by code. Confirm that NAICS-based Code is chosen.

   Based on your criteria, the ideal facilities are elementary and secondary schools. The NAICS-based code for these types of schools is 6111.

   Tip:

   To find the NAICS-based code for a business or industry type, you can look it up on NAICS & SIC Identification Tools.

4. For NAICS-based code, type 611.
5. For Search extent, leave it at Current map extent.

   

   On the map, make sure the full extent of Houston City, TX is visible.

6. Click Go.

   The results are added to the map, but there are results outside of Houston. You'll further limit the results by site.

7. Close the Legend window.
8. Under Refine your results, for By site, choose Houston City, TX and verify that the Name box is checked.

   

   The results are added to the map and listed in the search pane on the left side of the window. There are approximately 1,000 primary and secondary schools in Houston. In the pane, you can further refine the results to select the best candidates.

   Note:

   Demographic data is updated periodically so your results throughout the tutorial may differ.

9. Scroll down to the Number of employees parameter. For the minimum, type 125 and press Enter.

   

   The number of employees a facility has is a reasonably good indicator of how large the facility is. Only larger schools are good candidates, because shelters need to be able to hold large numbers of people.

   Now that you've narrowed the results, they no longer need to be clustered. You'll turn off clustering so each school gets its own pin on the map.

10. In the Points of Interest Search pane, click Next.
11. In the Style search results step, uncheck the Cluster points box.

    

    Each of the schools is drawn on the map. However, these results still include points in both the evacuation zone and the floodplain (the colorful buffer zones). Before you save the layer, you'll remove facilities in these areas.

12. If necessary, close the Legend window.
13. On the toolbar, click the Select button.

    

14. On the map, draw a rectangle over the points in the evacuation zone. (You may want to zoom in to the evacuation zone so you can get a more accurate bounding box.)

    

    The points in the box, as well as the Houston city boundary, are selected. Pop-ups appear for the points and the boundary.

15. Close the Houston City, TX pop-up. In the other pop-up, click Remove these points.

    

    The selected schools are removed from the map.

    

    Depending on how you drew your bounding rectangle, you may still need to remove points.

16. If any points remain inside the evacuation zone, select and remove them.

    Tip:

    To select a single point, you can click the pin rather than drawing a bounding rectangle over it.

    Next, you'll switch layers and use the same process to remove the schools that lie within floodplains.

17. On the ribbon, click the Emergency Shelter Locations tab.
18. In the project pane, turn on the USA Flood Hazard Areas layer and turn off the Harris_County_Hurricane_Evacuation_Zones layer.
19. If necessary, close the Legend window.
20. Zoom to Houston.
21. Select and remove all the schools within floodplains.

    Because these polygons are much smaller than the evacuation zones, you may have to zoom in to individual points to determine if they need to be removed. If any additional information windows appear, you can close them.

    All the schools remaining on the map satisfy the following criteria:

    • Must have at least 125 employees
    • Must be outside Hurricane Evacuation Zones
    • Must not overlap with USA Flood Hazard Areas

    You'll use the remaining points in the site-selection process. Before saving this result, you'll turn the school points into sites by drawing 2-mile buffers around them. Because site selection uses demographic data to make decisions, you'll enrich these points. Each site will receive the values of the demographic variables within the buffered zone.

22. In the project pane, turn off the USA Flood Hazard Areas layer. Click Close.
23. Draw a rectangle around the remaining points on the map to select them. In the Selection window, click Create buffers.

    

24. On the Rings tab, for Radius, change the first value to 2. Delete the other values and confirm that miles is chosen as the unit of measurement.

    

25. Click Apply.

    

    The buffers are created. Next, you'll create sites and save them as a new layer.

26. In the pop-up, click Create sites. Click Create 85 sites (your number of sites may differ).

    The Save Created Sites window appears.

27. For Create a new layer, name the layer Schools Outside Flood Zones and add your name or initials.

    

28. Click Save. Close the pop-up.

    A new layer is added to your project pane below Point Locations (Sites).

    

29. On the toolbar, click the Clear map button.

    

30. In the Clear Map window, expand Sites and check the box for Houston City, TX and click Clear.

    The yellow boundary line is removed from the map.

You've created a project in Business Analyst. You also added data to the map using different methods, including linking ArcGIS Online content, importing a .xlsx file, and searching by NAICS code. Using this data, you began to narrow down some of the schools in Houston that are potential shelter locations and saved them as sites. The analysis you performed was based on characteristics of the schools' locations. Next, you'll perform a suitability analysis to rank the schools based on the demographics of the populations they'd serve.

Perform a suitability analysis

Previously, you created a project in Business Analyst. Then, you added relevant data to determine what schools in the city of Houston could potentially be suitable hurricane shelters. Next, you'll use the school sites you created to run a suitability analysis. A suitability analysis applies chosen demographic data to each site using the population inside the buffer rings. First, you'll map the layer of existing hurricane shelters and add data for poverty, population density, income, and age. Then, you'll create weights for each of these variables to determine how much they'll affect the location of your new shelter.

Add demographics variables

Using the school sites you created, you'll add data that will help you determine the suitability of each potential school. First, you'll map existing hurricane shelters to make sure none of the new sites are too close. Then, using the demographic data browser, you'll enrich the school sites with population density, poverty, income, and age data.

1. If necessary, sign in to Business Analyst and open your Emergency Shelter Locations project.
2. On the ribbon, click Run Analysis and choose Suitability Analysis.

   

   The Suitability Analysis pane opens.

3. If necessary, click Get Started.

   You have two options: to add sites from the project or to start with features on the map. This is where the sites saved to the Point Locations layer can be used in future analyses. For now, the points are on the map.

4. Click Start with Features on Map.

   

   A window appears, asking you to confirm the later you want to use in the analysis.

5. In the Select Sites or a Layer window, confirm that Sites is chosen and click Apply.

   

   The Suitability Analysis pane updates to list the sites.

6. In the Suitability Analysis pane, click Next.
7. For Click on a criteria list to run analysis, click Add Criteria and choose Add point layer.

   

   The Add Point Layer window opens.

8. Check the houston-shelters box. Close the Add Point Layer window.

   The houston-shelters.csv file is added to the analysis. The location of existing shelters is automatically weighted inversely, or negatively. The closer a potential new shelter site is to existing shelters, the less it will be considered. The preliminary ranked results of the location criteria are shown at the bottom of the screen. Next, you'll add demographic criteria.

9. In the Suitability Analysis pane, click Add criteria and choose Add variables from data browser.

   

   The Data Browser window appears. By clicking a variable category, you can add Esri Demographics data to the analysis. You want to find vulnerable populations that will be the most difficult to evacuate. Age is a good indicator; people over the age of 65 are usually not as mobile, whether because of health problems or the loss of a driver's license, for example. Because of this, shelter sites near high populations of the elderly would be more useful.

10. In the Data Browser window, click Age.

    

    The Age category is displayed, showing the three most popular variables associated with age. These variables show mean and median age as well as total population over the age of 18.

11. Under Other Options, click Show all 'Age' variables.

    

    There are more than 2,400 possible variables to choose from. While these are all helpful variables, you don't need to know the ages of the entire population, just the number of people over 65 years old.

12. Click Create a custom variable.

    

    The Create Custom Variable window appears.

13. In the Create Custom Variable window, on the Age variable tab, for Set age range, move the age slider on the left to 65.

    

    Leave the other settings unchanged. The year will automatically be the most current for which data is available, and gender will include both males and females.

14. Click Add. In the Save Custom Variable window, click Save.

    The new variable is saved to your data browser and selected by default.

15. Check 2022 Both Ages 65+.

    

16. At the top of the window, click Categories.

    

    You return to the list of categories. The next variable you want to consider in your analysis is poverty. If a family can't afford to evacuate, they may need to go to a shelter.

17. Click the arrow on the right. On the second page, click Poverty.

    

18. For Popular Variables, select the percent button next to 2020 Households Below the Poverty Level (ACS 5-Yr). Then, check the variable's box.

    

    Percentage is more indicative of a neighborhood's need. The percentage considers the area's poor households in relation to the total population. Like the age variable, Households Below the Poverty Level is added to Selected variables.

19. Click Categories and click the Key Facts category (If necessary, click Show all 'Key Facts' variables under Other Options). Add the following variables by doing the following:
    • Expand 2022 Key Demographic Indicators (Esri) and check 2022 Total Population (Esri).
    • Expand 2022 Income (Esri) and check 2022 Median Household Income (Esri).
20. Click Selected variables. Confirm that all four variables are in the menu.

    

21. In the Data Browser window, click Apply.

    The variables open in the Suitability Analysis pane. Because there are five variables (including the shelters), each is automatically assigned an equal weight of 20 percent.

    

    Next, you'll change the weights to reflect the relative importance of your criteria.

Weight variables and location

Creating weights for variables allows you to choose certain variables that you want to emphasize more or less in your decision making. Because you're trying to site a shelter that will serve those who are not easily able to evacuate, you'll give age and poverty greater weights. There is also the option to give variables inverse weights, as you did for the current shelter sites. This option makes them count negatively in the suitability analysis.

1. In the Suitability Analysis pane, for the 2022 Both Ages 65+ variable, drag the Weight slider up to 25 percent. Click the lock icon.

   

   All the other variables' weights are automatically adjusted. The weights must always add up to 100.

2. Change the weight for 2020 HHs: Inc Below Poverty Level (ACS 5-Yr) to 30 percent. Click the lock icon.
3. For 2022 Median Household Income (you may have to scroll down to see it), click More options.
4. For Influence, choose Inverse. Change the weight to 10 percent.

   

   Unlike the other variables, a higher median income indicates less need for a nearby shelter site. People with higher incomes tend to be able to afford to travel farther away during an evacuation. While you'll still include these populations in your analysis, you'll account for the greater need elsewhere.

5. Under Refine results, click the Filter button.

   

6. Change the limit on the left to 0.6 by dragging the bar or by typing 0.6 into the box. (The right limit may change on its own.)

   

   Note:

   Demographic data is updated periodically. Your results may differ.

7. Click Done.

   The results of your suitability analysis will have a score of 0.6 or above, where 1 is perfectly consistent with all criteria. Scores are calculated using the percent difference of the value for a given site compared to the target value you selected.

8. Click Export and choose Export to a new suitability layer.
9. Name the new layer Suitable Shelter Locations and add your name or initials.

   

10. Click OK.

    The new layer is saved to the project.

11. On the ribbon, click the Emergency Shelter Locations button.
12. Turn off the houston-shelters layer.

    

    You can point to the circles and read the pop-up information. The empty circles represent schools that did not meet your minimum final score of 0.6; the colored circles represent the schools that did. Your final analysis may vary slightly depending on how many schools you converted to sites.

    Note:

    To reopen or edit the suitability analysis later, click the ellipsis next to the layer and choose Open analysis. There, you'll be able to adjust the criteria, weights, and other settings to modify the analysis. If you have previously filtered the results, you'll only be able to see the results of that selection.

In this tutorial, you ran a site suitability analysis that identified the top locations for shelters based on need. These sites fit most of the humanitarian requirements for hurricane shelters. Structural compliance will still need to be evaluated, but it is much more feasible to look at the remaining sites than the 900 you started with.

Site selection, with the capability to weight variables both positively and inversely, is a powerful tool that can be used to benefit communities and businesses alike. Depending on what variables you choose to include, you can evaluate everything from insurance ownership to fast food habits.

You can find more tutorials in the tutorial gallery.

Acknowledgements

• Demographic data accessed with ArcGIS Business Analyst. Terms of use.