Calculating Kernel Densities
Included in this tutorial:
Accessing the Kernel Density tool
Calculating a Kernel Density raster from point feature locations: Setting parameters and examining example outputs
Calculating a Kernel Density raster with values from the input feature attribute table
Software version in examples: 3.0.0
Tutorial Data: The tutorial includes demonstration with sample data available here.
Credits: L. Meisterlin and Nikolas Michael (2022)
This tutorial demonstrates using the Kernel Density tool in ArcGIS Pro to calculate kernel density rasters.
The demonstration includes three examples: two examples generated by calculating the density of the input location of points within a feature class, examining changes to the input parameters; and a third example generated from quantities within the feature class’s attribute table.
Related tutorials: This tool requires either the Spatial Analyst Extension in ArcGIS Pro. To verify that you have the Network Analyst Extension, see Which Esri Extensions Is My Software Licensed to Use?
Accessing the Kernel Density Tool
Find the Kernel Density tool in the Geoprocessing Pane by clicking through Spatial Analyst Tools > Density > Kernel Density. You can also find it by using the Geoprocessing search bar.
Calculating a Kernel Density Raster from Point Locations
Setting the Parameters
The Kernel Density tool’s parameters include the standard input and output information as well as several options that determine how the density calculations are performed and the units represented by the output values.
Geoprocessing panel with the parameters for Example 1 (below)
Input point or polyline features: Point or polyline feature classes, from which you will calculate the kernel density raster.
Because polygons features are not valid inputs for this tool, see the Creating Points from Polygons tutorial as well.
Population field: You can use this to ask the tool to calculate density of a certain feature, this is useful if you are using points representing, for example, housing units and want to control for how many people live there. In this screenshot, we only want the density of the points, so we have chosen to leave the Population field as NONE.
Output Raster: The location and name of the output kernel density raster.
Output cell size: Specify the output cell size of the kernel density raster. The cell size is determined by this value squared, in the units of the input dataset’s Coordinate Reference System (CRS).
In our examples, setting this parameter to 700 generates the output’s cell resolution as 700 meters x 700 meters (because the input data’s CRS unit is meters).
Search Radius: Specify the kernel density search radius. This is the distance from each output cell that defines the area in which the density value (for that cell) is calculated. Again, this value is specified in the units of the input data CRS.
Area Units: The area units of the output density values which is set, by default, to the units of the input dataset’s CRS.
Output Cell Values: This specifies what the values in the output raster represent, either densities (quantity per unit area for each cell) and expected counts (the expected count of points per cell based on the density calculated surrounding the cell.)
For more information on interpreting the output values (and how expected counts are calculated from densities), see this article on the Esri ArcGIS blog.
Method: Specifies whether planar or geodesic method will be used. Planar is the default, and most useful at a larger scale, and supports both points and polylines. The geodesic method only supports points and is most useful when working at a small scale (large area).
Input Barrier Features: A dataset that uses polygons or polylines to increase the distance between a point and a cell when calculating density.
To execute the tool, click Run.
Examining Example Outputs
Example 1 (below, left) was produced from the sample points with a 700-meter cell size and a 3000-meter search radius.
Example 1 output
Example 2 (below, right) was produced from the sample points with a 70-meter cell size and a 300-meter search radius.
Example 2 output
Calculating Kernel Density Rasters with Values from the Input Attribute Table
Example 3 (below) was produced from sample points with a 700-meter cell size and a 3000-meter search radius, the same as Example 1.
However, Example 3 was created with values from the input attribute table assigned to the Population Field parameter. As a result, the density of the values at each point feature’s location is calculated.
Example 3 output