Find patterns in fire incidents

A city can have thousands of fire incidents over a period of years. With so much data, simple visualization may not be enough to determine where fire risk is higher. In this tutorial, you'll perform statistical analysis on fire incident data to find spatial patterns that might otherwise be invisible. Specifically, you'll pinpoint statistically significant clusters of incidents. These clusters are known as hot spots.

First, you'll add a layer of fire incidents to a web map in ArcGIS Online. You'll filter and style the data to better understand what it shows. Then, you'll perform hot spot analysis to find significant clusters of fire incidents, which may indicate areas where fire risk is higher.

This tutorial was last tested on September 25, 2024.

Requirements

• ArcGIS organizational account with a Publisher, Facilitator, or Administrator role (see options for software access)

Find patterns in fire incidents

Map the data

First, you'll add the fire incident data to a web map in ArcGIS Online.

1. Sign in to your ArcGIS organizational account.

   Note:

   If you don't have an organizational account, see options for software access.

2. On the ribbon, click the Map tab.

   

   A new, untitled web map appears in Map Viewer. The map displays a default basemap, which provides contextual information about landmasses, water bodies, national borders, and so on.

   

   The Layers pane lists layers, or collections of spatial data, that appear on the map. Currently, the pane is empty. You'll add a layer that contains data about fire incidents in Naperville, Illinois, between 2017 and 2021.

3. In the Layers pane, click Add.

   

   Tip:

   If you don't see the Layers pane, you can make it appear by clicking the Layers button on the Contents (dark) toolbar.

   By default, the Add layer pane shows content in your account. You can also add data from other sources, such as your organization. The layer you'll add is a publicly shared layer hosted in a different organization.

4. Click My content and choose ArcGIS Online.

   

   Now, when you search for layers to add to the map, you'll search all public content on ArcGIS Online.

5. In the search bar, type Naperville Fire owner:Learn_ArcGIS and press Enter.

   Tip:

   Adding owner: allows you to search for content from a specific account.

6. In the list of results, for Naperville 2017-2021 Fire Incident Calls, click Add.

   

   The layer is added to the map. The map's extent changes to zoom to the data, which is focused around Naperville, Illinois.

   

   Note:

   If you completed the previous tutorial in this series, Geocode fire data, you published a layer of fire data from Microsoft Excel to your ArcGIS account. It's possible to add that data to your map, but for the purposes of this tutorial, you'll use a more extensive dataset than the one in that tutorial. You can follow the steps in the previous tutorial and this one to publish your own fire incident data from Excel and add it to a map in ArcGIS Online.

Filter the data

The layer you added contains all incidents where the Naperville Fire Department was called. However, not all of these incidents correspond to actual fires; some are for medical emergencies, accidents, or false alarms.

Your goal is to learn where fires are occurring to better understand fire risk, so you're not interested in calls that aren't related to fires. You'll investigate the data more closely and create a filter to hide incidents that aren't relevant to your goal.

1. In the Add layer pane, click the Back button.

   

   You return to the Layers pane. It now lists one layer: Naperville 2017-2021 Fire Incident Calls.

2. For Naperville 2017-2021 Fire Incident Calls, click the Options button.

   

3. In the list of options, click Show table.

   

   A table appears at the bottom of the map. The table contains additional information, or attributes, associated with the layer. The rows in the table correspond to individual points on the map (features), while the columns correspond to types of information (fields).

   The table also lists the total number of features in the layer. This layer has 43,161 features.

   Note:

   If you look at the table of a layer you published from Excel, the table will correspond to the spreadsheet table you used to make the layer.

4. Scroll through the table until you find the Incident Type Code (National) field.

   Tip:

   If a column is too small to see the entire heading, you can expand it by dragging its right edge to the side.

   

   This field refers to the incident's National Fire Incident Reporting Systems (NFIRS) incident type code. Generally, codes between 100 and 199 have to do with fires, while other codes involve other types of incidents.

   Note:

   See the NFIRS Complete Reference Guide for more information about the codes, including a full list of what they refer to.

5. Close the table.

   You'll filter the layer to only show incident calls related to fires.

6. On the Settings (light) toolbar, click the Filter button.

   

   Note:

   If the Filter button is grayed out or unavailable, make sure that Naperville 2017-2021 Fire Incident Calls is selected in the Layers pane.

   The Filter pane appears. Filters are added by creating a condition that indicates what data in your layer you want to show, based on information in the table. Your condition will create a filter so only fire incident calls with codes between 100 and 199 are shown.

7. In the Filter pane, click Add new.
8. Under Condition, click Object ID and choose Incident Type Code (National).
9. Click is at least and choose is between.
10. Change the first number to 100 and the second number to 199.

    

    This condition limits the layer to features with an incident type code between 100 and 199. A histogram under the condition shows how many features have each code. By filtering your data this way, you hide a considerable number of features that are not related to fire incidents.

11. Click Save.

    The layer is filtered. The map now displays far fewer features, but all features shown are directly related to incidents involving fire.

    

    Tip:

    If you ever want to remove the filter and view the full dataset again, reopen the Filter pane and click Remove filter.

Change the style

Next, you'll change the layer's style, or the way it appears on the map. Currently, every fire incident point has the same appearance: a green circle. However, you can style the data in other ways, including based on the information in the table.

Your data contains fire incidents from 2017 to 2021. You'll style the data so incidents from different years have a different appearance. This way, you'll be able to visually inspect trends in fire incident locations over time.

1. On the Settings toolbar, click Styles.

   

   The Styles pane appears. First, you'll choose the field from the table you want to map. You'll choose a field that indicates the year.

2. In the Styles pane, under Choose attributes, click Field.

   

3. In the Select fields pane, in the search bar, type Year. In the list of results, select Alarm Date - Year.

   

4. Click Add.

   The style on the map changes automatically, showing incidents from each year with a different symbol. By default, the style is based on size, which isn't helpful in this case. You'll change the style to give each year a unique symbol.

5. In the Styles pane, under Pick a style, scroll to the bottom and click Types (unique symbols).

   

   You'll customize the symbols.

6. Under Types (unique symbols), click Style options.

   

   More style options appear. Each year has been assigned a unique color, but the symbols are small. You'll increase their size.

7. Click Symbol style.

   

8. In the Symbol style window, set Size to 10 px.

   

9. Close the Symbol style pane.

   The changes have automatically been applied to the map.

   

   From a visual examination, there doesn't appear to be a pattern connecting the spatial location of fires and the year they occurred; fires are evenly distributed across Naperville each year.

   However, there do appear to be more fires that occurred in 2017 (red) than any other year. The Style options pane lists the number of features per year, which confirms that 2017 had over twice the number of fire incidents (196) as any other year (the next closest is 2020, with 82).

   

   Why does 2017 have more incidents? It might be because fire risk mitigation efforts since then have successfully reduced the number of fires. Alternatively, such a drastic and immediate reduction might suggest issues with the completeness of the data. (This tutorial uses only a subset of the total Naperville fire incident calls dataset.) You should always think critically about what your data is showing and ensure its accuracy.

   The years are currently out of order. They also have comma separators, which are not appropriate for years. As the years also appear in the map legend, you'll edit them appropriately.

10. Click 2,017.

    

    The name becomes editable.

11. Change the name to 2017 (with no comma) and press Enter.
12. Rename the other years to remove the commas.
13. Drag the handle next to 2020 to position it after 2019 and before 2021.

    

    The years now appear in the correct order.

    

    The change is also reflected in the legend on the map. You've completed your changes to the style.

14. At the bottom of the Style options pane, click Done. At the bottom of the Styles pane, click Done.

Find hot spots

Visualizing fire incidents by year didn't lead to any insights about spatial patterns that can help you better understand fire risk in the area. Next, you'll perform analysis on the data to calculate statistically significant clusters of fire incidents. This type of analysis is called hot spot analysis.

Hot spot analysis is statistical in nature, and thus less subjective than simply visualizing data on the map. It will inform you with a high level of statistical certainty where fire incidents are significantly clustered within Naperville.

1. On the Settings toolbar, click Analysis.

   

2. In the Analysis pane, click Tools.

   

3. In the search bar, type Hot Spots. In the list of results, click Find Hot Spots.

   

   The Find Hot Spots tool requires several parameters to run. The first parameter is the input layer (the layer you want to run the tool on).

4. For Input layer, click Layer and choose Naperville 2017-2021 Fire Incident Calls.

   The layer is added as the input.

   

   The listed count of features is 487, which is the number of fire incidents, rather than 43,161, which is the total number of features in the dataset. When tools are run on a filtered dataset, only the features shown by the filter are used for analysis.

   Next, you'll set parameters for the hot spots themselves. Hot spot analysis aggregates point data into polygon bins and uses the number of points in each bin to determine clusters. By default, the polygon bins are a grid of squares. You'll change it to a grid of hexagons, because hexagons can better represent curves in the data.

   Note:

   To learn more about the difference between square and hexagon aggregation, see Why hexagons?

5. For Aggregation shape type, choose Hexagon cells.

   

   Additionally, you could set parameters to perform hot spot analysis based on a specific attribute field in your data or define the exact size of the hexagon bins used for aggregation. For this tutorial, you only want to determine hot spots based on the number of fire incidents, and the default bin size is fine, so you won't change more parameters.

6. For Output name, type Fire Incident Hot Spots and add your name or initials to ensure the layer name is unique within your organization.

   Note:

   New items created by analysis operations must have unique names within your organization; otherwise, their URLs will conflict. Once the layer has been created, you can rename it in your map.

   

   Analysis tools perform automatic statistical calculations on features, but cost credits depending on the type of analysis and the number of features being analyzed.

7. Under Environment settings, click Estimate credits.

   The estimated credits needed to run the tool is 0.487, a very small amount.

8. Click Run.

   The tool runs. After about a minute, the analysis is completed and the result layer is added to the map. You'll turn off the original fire incidents layer to better see the results of your analysis.

9. In the Layers pane, point to Naperville 2017-2021 Fire Incident Calls and click the Visibility button.

   

   The layer is hidden. You can show it again by clicking the Visibility button again. The map now only shows the hot spots layer.

10. On the map, zoom in using the mouse scroll button. If necessary, pan the map by dragging it.

    

    The map indicates a clear group of hot spots in the northwest corner of Naperville. This area represents a statistically significant cluster of fire incidents. The white hexagons indicate areas where a relatively average number of fire incidents occur.

    While fire incidents occur all throughout Naperville, the significant clusters in the northwest indicate an area where fire risk is particularly high compared to the rest of the city. It might be a good idea to be especially prepared to respond to fires in this area.

Save the map

Before you continue, you'll make some final changes to the map, such as adjusting the items in the Layers pane and changing the basemap. Lastly, you'll save your map, so you can access it again from your account's content.

1. In the Layers pane, for Naperville 2017-2021 Fire Incident Calls, click the Options button and choose Rename.

   

2. Rename the layer to Fire Incidents 2017-2021 and click OK.

   Now, the layer's name reflects that it shows only fire incidents, not all calls. You'll also remove your name or initials from the hot spots layer. The original hot spots layer will still have your name or initials in it, they just won't show up in the map.

3. Rename Fire Incident Hot Spots Your Name to remove your name or initials from the end of the layer name.

   You'll rearrange the order of the layers so the incidents appear above the hot spots. Generally, it is recommended that you put point layers (like the incidents) above polygon layers (like the hot spots).

4. In the Layers pane, drag Fire Incidents 2017-2021 above Fire Incident Hot Spots.

   

   Because your hot spots layer has many hexagons with white coloring, you'll change the basemap to a darker basemap so they stand out more clearly.

5. On the Contents (dark) toolbar, click Basemap.

   

6. Scroll through the list of basemaps and choose Dark Gray Canvas.

   Now, the hot spots layer is more visible.

   

   Your map is finished, so you'll save it.

7. On the Contents toolbar, click Save and open and choose Save as.

   

   When saving, it's recommended that you add appropriate metadata so users know what your map is for.

8. In the Save map window, for Title, type Naperville Fire Incidents 2017-2021.
9. For Tags, type Firefighting and Hot Spots, pressing Enter between each tag.
10. For Summary, type Hot spots of fire incidents in Naperville, Illinois, between 2017 and 2021.

    

11. Click Save.

    Your map is saved. Now, you can access it from your account. Before you finish, you'll share the map so others can view it.

12. On the Contents toolbar, click Share map.

    

13. In the Share window, for Set sharing level, choose Everyone (public). Click Save.

    You also must share the layers used in the map.

14. In the Item sharing successfully updated window, click Review sharing. In the Review sharing window, click Update sharing.

    Now, anyone with the URL can view your map.

In this tutorial, you mapped fire incident data in Naperville, Illinois. First, you added a layer of fire incident calls to a map. Then, you filtered the data to only show calls directly related to fire. You changed the layer's style to show fires by year and performed hot spot analysis to determine spatial patterns in the data.

This workflow can be completed with your own fire incident data. If you want an extra challenge, try to filter your own data, and then change the style and find hot spots. What spatial patterns can you find in your data?

You can find more tutorials in the tutorial gallery.

Acknowledgements

• The Naperville 2017-2021 Fire Incident Calls feature class was compiled using Annual NFIRS Public Data managed by the Federal Emergency Management Agency (FEMA). Retrieved on October 26, 2023. This product uses the FEMA OpenFEMA API but is not endorsed by FEMA. The federal government or FEMA cannot vouch for the data or analyses derived from this data after the data has been retrieved from the agency's websites. See OpenFEMA Terms and Conditions for more information. Data on FEMA.gov is free of copyright and may be copied and distributed without permission.
• Topographic map sources: Esri, TomTom, Garmin, FAO, NOAA, USGS, OpenStreetMap contributors, and the GIS User Community
• Dark Gray Canvas map sources: Esri, TomTom, Garmin, FAO, NOAA, USGS, OpenStreetMap contributors, and the GIS User Community