Track vehicle behavior to apprehend a suspect

Classifying movement events is an analytical technique that identifies how people and things move through an environment. Determining where individuals make turns or change speed provides insights into movement patterns that can be used for actionable responses.

In this tutorial, you're a law enforcement officer planning a raid on a suspected criminal while they move from one area to another in a vehicle. To ensure the raid is successful, you'll analyze vehicle movement behavior to determine when the suspect's vehicle is most likely to slow down or stop. You'll perform the analysis in ArcGIS AllSource, intelligence analysis software for investigative, geospatial and link analysis.

This tutorial was last tested on July 19, 2023, using ArcGIS AllSource 1.0. If you're using a different version of ArcGIS AllSource, you may encounter different functionality and results.

Requirements

• ArcGIS AllSource (learn more)

Track vehicle behavior to apprehend a suspect

Add the data

You'll start by reviewing a fictitious dataset of vehicle movement events at the area of interest. First, you'll download the data and add it to a project in ArcGIS AllSource.

Note:

All of the data used in this tutorial is fictitious and designed for educational use only.

1. Download the Camp Lemonnier Intelligence Data.

   A zipped folder named Camp_Lemonnier_Intelligence.gdb is downloaded to your computer. The .gdb extension means the folder contains a geodatabase, which is a folder format for storing geographic data.

2. Extract the zipped folder to a location you can easily remember, such as your Documents folder.

   Next, you'll create a project in ArcGIS AllSource and add the data to it.

3. Start ArcGIS AllSource. If prompted, sign in using your licensed ArcGIS account.

   Note:

   ArcGIS AllSource is an optional product that can be installed when you install ArcGIS Pro. To learn more about installing ArcGIS AllSource, go to the Install ArcGIS Pro documentation page. If you don't have ArcGIS Pro or an ArcGIS organizational account, see options for software access.

   When you start ArcGIS AllSource, you're given the option to create a new project or open an existing one. If you've created a project before, you'll see a list of recent projects.

4. Under Blank Templates, click Map.

   

5. For Name, type Classify Movement Events Project. Click OK.

   The project, which will contain all the maps and data for this workflow, is created. Because you chose the Map template, the project includes a blank map.

   You'll add data about vehicle movements from the geodatabase that you downloaded.

6. On the ribbon, click the Data tab. In the Import group, click the Add Data button.

   

   The Add Data window appears. You can add data from the project, your portal (ArcGIS Online), or your computer.

7. Under Computer, browse to the location of the extracted geodatabase that you downloaded. Double-click Camp_Lemonnier_Intelligence.gdb.

   

   The geodatabase contains three feature datasets: Administrative_Data, Cell_Phone_Data, and Vehicle_Data. You're interested in tracking vehicle behavior, so you'll add the vehicle dataset.

8. Double-click Vehicle_Data.

   

   A feature dataset contains multiple feature classes. Feature classes are collections of geographic features (such as points, lines, or polygons) that can be added to a map. The Vehicle_Data feature dataset has 11 feature classes, but you'll only use one, Vehicle_Data_All_Vehicle_Data, for this analysis.

9. Click Vehicle_Data_All_Vehicle_Data to select it.

   

10. Click OK.

    The feature class is added to the map. The map zooms to the extent of the data.

    

    Note:

    Your layer may have a different color than the example images.

    The layer includes a large number of points (more than 1 million) concentrated in and around the city of Djibouti in Djibouti, Africa. The area of interest includes an American military base named Camp Lemonnier.

    Each point represents the location of a vehicle at a certain point in time. Multiple points may correspond to the same vehicle as it moved throughout the day. You can learn more about a layer by opening its attribute table.

11. In the Contents pane, right-click All Vehicle Data and choose Attribute Table.

    

    The attribute table appears. Attributes are textual or numeric data associated with each feature. In the table, each row represents an individual feature, while each column represents an attribute field.

    

    This table includes fields describing the unique ID (OBJECTID), shape, latitude, longitude, and speed of each vehicle record. It also has a Date field, which has the date and time that the vehicle's location was captured, and a Track ID (Text) field, which contains an identifier for each unique vehicle. Because this data is fictitious, the track IDs do not correspond to any real-world vehicles.

    In the example image, the first seven vehicle records all belonged to the same vehicle, identified with a track ID of 0. These records were taken shortly after midnight on April 28, 2021, with each record taken one second after the previous one. (The first record has no time listed, meaning it was taken at 12:00:00 AM, the first minute of the day.) With these records, you can determine where a vehicle was located at a specific time and how fast it was going. By comparing multiple records belonging to the same vehicle, you can track a vehicle's movement patterns over time.

12. Close the table.

    Because there are more than a million vehicle records, looking at them all one by one is time-consuming. You can learn more about the data by creating a chart. You'll create a histogram chart to see the distribution of vehicle speeds across all of the records.

13. In the Contents pane, right-click the All Vehicle Data layer, point to Create Chart, and choose Histogram.

    

    The Chart Properties pane and the chart view appear. You'll choose a variable to chart.

14. In the Chart Properties pane, for Number, choose Speed.

    

    The chart view populates with a histogram showing the distribution of vehicle speeds in kilometers per hour (kph).

    

    The most common speeds for vehicles to travel are between 42 and 50 kph and 84 and 101 kph. The mean speed is 68 kph. This distribution of speeds suggests two types of roads: side streets where the speed limit is lower, and highways where the speed limit is higher. Relatively few vehicles travel at speeds between these two clusters of high-frequency speeds.

    While this chart gives you a general overview of vehicle behavior, with spatial analysis you can gain even more insight.

15. Close the chart and the Chart Properties pane.

Enable time settings

To classify movement events, your layer must be time-enabled. From exploring the attribute table, you know that your vehicle data has a time field. You can use this field to enable time settings for the layer.

1. In the Contents pane, double-click All Vehicle Data.

   The Layer Properties window appears. In this window, you can set many settings relating to the layer. First, you'll set whether the layer has a single time field or start and end time fields. Your data has only one time field.

2. Click the Time tab.
3. Under Filter using time, select Filter layer content based on attribute values.

   

4. For Layer Time, choose Each feature has a single time field.

   

   Next, you'll choose the time field from the list of attribute fields.

5. Confirm that Time Field is set to Date.

   Note:

   For a time field to be used to enable time settings, the field must follow certain rules. If you're performing this workflow with your own data and have trouble enabling time settings, try converting the time field into a date format.

6. Click OK.

   The layer is time-enabled. A timeline appears at the top of the map. When you point to the timeline, it shows the earliest and latest dates for the data.

   

Classify movement events

Now that your data is time-enabled, you're ready to classify movement events to better understand traffic patterns in the area. By classifying movement events, you'll be able to identify where vehicles turn and accelerate.

1. On the ribbon, click the Analysis tab. In the Tools group, in the Movement group, click Classify Movement Events.

   

   The Geoprocessing pane appears. The pane displays the Classify Movement Events tool. The tool requires several parameters. First, you'll choose the input dataset that you want to analyze and the dataset's unique ID field.

2. For Input Features, choose All Vehicle Data. For ID Field, choose Track ID (Text).

   Note:

   For an ID field to be valid, it must use a text data type, even if the ID uses numbers. This is why the vehicle dataset has two ID fields, Track ID and Track ID (Text). The first field uses a numeric data type, while the second uses a text data type. To learn more about field data types, read the ArcGIS field data types documentation page.

   Next, you'll choose the name of the output feature class that the tool will create to contain the results of the analysis.

3. For Output Feature Class, delete the text and type Vehicle_Movement_Events.

   

   Depending on your analytic workflow requirements, you may want to fill in several additional parameters. All of the following parameters are optional.

   • The Curvature parameter determines the number of points needed to classify a movement event as a turn event. If the event does not meet the number of points, it will be classified as a traveling event instead. The best number to use depends on the size of the objects you are measuring. The default value of 15 is appropriate for vehicles. Larger objects should use a higher value because it takes them more time to complete a turn.
   • The Number Of Points parameter determines the number of points evaluated before and after a given point when calculating the bearing difference. The best number to use depends on the speed of the objects you are measuring. The default value of 1 is appropriate when measuring the movement of pedestrians and vehicles. Faster objects, such as aircraft, should use a value of 5.
   • The Regions Of Interest parameter uses a polygon feature class to determine a specific area where movement events will be classified. This parameter is useful if you only want to analyze a subset of the data based on location. If you want to set regions of interest, you'll also need to set the Regions Of Interest ID Field parameter. Each region must have a unique ID field, similar to the input features.

   Because you're tracking vehicle movement events for the entire extent of the dataset, you don't need to change any of these optional parameters.

4. Click Run.

   Note:

   Because your dataset has a large number of features (more than 1 million), the tool may take several minutes to run.

   The tool runs. When it finishes, a notification appears at the bottom of the Geoprocessing pane.

   Tip:

   To learn more about the analysis, including how long the tool took to run, click View Details.

   Additionally, the Vehicle_Movement_Events layer is added to the map and Contents pane. It may be difficult to see the output layer because of the number of points in the original layer.

5. In the Contents pane, uncheck All Vehicle Data.

   

   Unchecking the layer hides it on the map. (You can always show the layer again by checking it.) Now, the map only shows the movement events.

   

   Even with the other layer turned off, it's difficult to gain much insight from the layer with its default appearance. Later, you'll change the layer's symbology to better see the movement events. Before that, you'll investigate the layer's attribute table.

6. In the Contents pane, right-click the Vehicle_Movement_Events layer and choose Attribute Table.

   The table appears.

   

   It contains a large number of fields, many of which are explained by the following list:

   • track_id—The unique identifier for the traveling feature. This ID is the same as the one you used as an input in the Classify Movement Events tool.
   • distance_diff—The distance in meters between each record and the record that preceded it.
   • time_diff—The time difference in seconds between each record and the record that preceded it. The first record for each unique traveling feature has the <Null> value for this and other fields, because no record preceded it.
   • speed—The speed of the traveling feature in meters per second based on the time and distance elapsed since the previous record.
   • speed_mph—The speed in miles per hour.
   • speed_kph—The speed in kilometers per hour.
   • acc_event—A description of the acceleration event, or how the traveling object's speed was changing during each record. Objects can be traveling (no change in speed), accelerating (speed is increasing), or decelerating (speed is decreasing/stopping).
   • turn_event—A description of the turn event, or how the traveling object's direction was changing during each record. Objects can be traveling (no change in direction), stopped, doing a left turn, or doing a right turn.

   The key fields created by the tool are the acc_event and turn_event fields, which describe the movement events of each vehicle at each point in time.

7. Close the table.

Filter the results

Your analysis results cover a wide area. However, you're most interested in tracking the movement patterns of a single vehicle that belongs to a suspected criminal. Before you continue, you'll filter the results to show only movement events belonging to that vehicle, which has a track ID of 743.

1. In the Contents pane, right-click Vehicle_Movement_Events and choose Build Definition Query.

   A definition query is an expression that filters a dataset to show only a subset of the data. You can filter data based on attributes from the table.

2. Click New definition query.

   

   You know the track ID of the vehicle you're interested in, so you'll filter the dataset based on the track_id field.

3. For Query 1, create the expression Where track_id is equal to 743.

   

4. Click Apply. Click OK.

   On the map, the dataset is filtered. Now, only records with a track ID of 743 are shown.

   

   All of these records belong to the same vehicle, the vehicle of your suspect. Based on the map, the suspect primarily traveled along a highway from the sparsely populated western mountain region to the populated eastern urban area.

   Based on other intelligence gathered about the suspect, you suspect that they take this route frequently to travel from one base of operations to another. By looking at the classified movement events of their vehicle on this route, you can predict their future behavior.

Change the symbology

To better understand the movement events, you'll change the way the layer looks, also known as its symbology. You can symbolize a layer based on data in its attribute table. You'll symbolize it based on the acc_event field. The result will show where vehicles accelerated and braked on the map, which will help law enforcement officials pinpoint the best areas to perform a raid on a moving vehicle.

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

   The Symbology pane appears. You want to change the symbology to use the unique values of an attribute table field.

2. In the Symbology pane, for Primary symbology, choose Unique Values.

   

   Next, you'll choose the field on which to base the symbology.

3. For Field 1, choose acc_event.

   

   The bottom of the pane populates with all of the unique attributes for this field. There are seven attributes. While you can change every attribute's symbol individually, the attributes you're most interested in are those involving braking, as it is easier to perform a raid on a moving vehicle when it slows down.

4. In the list of classes, for the Decelerating symbol class, click the Format symbol button.

   

   A gallery of symbol types appears. You'll choose a red symbol, which suggests slowing down.

5. In the gallery, click the Circle 3 symbol.

   

   On the map, all decelerating/braking movement events are updated with the circle symbol.

   

   There are decelerating events distributed throughout the route. On the surface streets in the city, braking is generally predictable and corresponds to stoplights or other traffic signs. On long highways, however, such as the one leading out of the city, there are fewer traffic signs; here, braking may correspond to changes in elevation or turns.

   For the purpose of this tutorial, assume that it's preferable for an arrest to be made on the vehicle when it's outside the city. The suspected criminal may be dangerous, and attempting to arrest them in a populated area could lead to collateral damage.

6. Zoom to the far west decelerating events.

   

   Tip:

   To zoom to a specific area of the map, press Shift and draw a box around the area.

   This area has a large number of decelerating events across a relatively short distance, possibly due to the mountainous terrain of the area (which is visible on the basemap). How fast is the vehicle traveling across this area? You can open the pop-ups of some of the decelerating events to find out.

7. On the map, click one of the braking events to open its pop-up.

   

   According to the pop-up, even though the car is slowing down, its speed is still about 100 kph (or 60 mph). The vehicle is still traveling fast, so this may not be the best place to plan a raid.

8. Close the pop-up.

   Next, you'll look at an area that is closer to the city, but still far from a densely-populated area.

9. Zoom back out to the full extent of the data. Zoom to the decelerating events at the western edge of the city.

   

   A group of five consecutive decelerating events occur as the road enters a roundabout.

   

   It seems likely that the vehicle slowed considerably before it entered the roundabout.

10. Open the pop-ups of each of the five decelerating events, noting the speed during each.

    The westmost decelerating event, which occurred first, happened at a speed of 100 kph or 60 mph, the same speed the vehicle was traveling earlier on the highway. However, the eastmost decelerating event has a speed of about 60 kph or 40 mph. Because each event is recorded one second after the previous, that means that in a span of 5 seconds, the vehicle slowed to about two-thirds its previous speed.

    This area may be a place to consider for the raid. The area is on the outskirts of the city, so the civilian presence may be relatively light, while the vehicle is forced to slow itself considerably to proceed. It might be a good idea to scout the location on the ground to further assess the possibility of conducting the raid there.

11. Close any open pop-ups.

    You'll bookmark this location so you can return to it quickly in the future.

12. On the ribbon, click the Map tab. In the Navigate group, click the Bookmarks button and choose New Bookmark.

    

13. In the Create Bookmark window, for Name, type Roundabout. Click OK.

    Now, if you click the Bookmarks button, you can choose this bookmark and immediately navigate to this area of the map.

14. Return to the full extent of the data. On the Quick Access Toolbar, click the Save button.

    

    The project is saved.

In this tutorial, you used ArcGIS AllSource to classify the movement events of vehicles. Using these classified movement events, you were able to track the vehicle behavior of a vehicle belonging to a suspected criminal, and could identify a location where an arrest might be made with relatively high odds of success and safety. In the real world, movement event analysis can be performed on a variety of datasets, not just vehicle records.

The dataset you downloaded at the beginning of the tutorial includes the All_Cell_Phone_Data feature class (contained in the Cell_Phone_Data feature dataset). It contains cell phone locations over time. You could run this same analysis on the cell phone data to identify areas where a suspect is on foot and track their movement patterns as a pedestrian. If you want an optional challenge, try running this workflow on that dataset.

This workflow, in conjunction with other intelligence workflows that can be done in ArcGIS AllSource, can be used by law enforcement and military personnel to help track down offenders or insurgents.

You can find more tutorials in the tutorial gallery.

Acknowledgements

• The Camp Lemonnier Intelligence Data is fictitious sample data created for this tutorial by David Attaway, PhD, Esri.
• World Hillshade map: Esri, Airbus DS, USGS, NGA, NASA, CGIAR, N Robinson, NCEAS, NLS, OS, NMA, Geodatastyrelsen, Rijkswaterstaat, GSA, Geoland, FEMA, Intermap, and the GIS user community
• World Topographic map: Esri, HERE, Garmin, FAO, NOAA, USGS, © OpenStreetMap contributors, and the GIS User Community