EPIC - Custom GeoWave Heatmap for Spatial Binning Aggregations and Statistics

[mzagorskirs]

Purpose/summary of Epic: Get up-to-speed with GeoServer by way of examples and hands-on development, which will work toward a goal of creating an aggregation heatmap in GeoWave.

Tasks:

• 1: Spike/learning experience: Develop a render transformation GeoServer plugin. The plugin should be able to translate the data (e.g. from vector <-> raster, or vector -> vector) and shift the data by a specified amount of degrees. Data could be a shapefile, but probably best to use the GDELT vector data. Code should be developed within a personal repo as a plugin for GeoServer using Java. Through this task, the developer will get up-to-speed with creating GeoServer plugins and creating render transforms.
• 2: Look at GeoServer example of heatmap and get it working using GDELT data, with the ultimate goal of writing own heatmap plugin in JAVA.
• 3 - 5: The ultimate goal: the heatmap plugin that GeoServer uses queries all the data in GeoWave, however, we want to use GeoWave's aggregations (and eventually statistics) in the heatmap query.
• 6: Develop GeoWave customization of the heapmap process using spatial binning statistics.
• 7 - 8: User preferences and automations.
• 9: Add spatial constraints to query builders to minimize processing time.
• 10: Configurations to handle tiling in GeoServer.
• 11: Convert to WGS84.
• 12: Finalize integration tests.

Definition of Done: This GitHub issue will be considered done when these items have been met:

• Task 1 - GeoServer plugin that shifts vector data by a specified amount of degrees.
• Task 2 - Local run of GeoServer heatmap example using GDELT data.
• Task 3 - Develop an Integration Test that will test the application of the Heatmap SLD to some created unit test data.
• Task 4 - Port over the GeoTools version of the heatmap process and research spatial binning methods in GeoWave.
• Task 5 - Develop GeoWave version/customization of the heatmap process using spatial binning aggregation: field sum and field count.
• Task 6 - Develop GeoWave customization of the heapmap process using spatial binning statistics.
• Task 7 - Add more parameters to SLD for user preferences.
• Task 8 - Add a dynamic GeoHash algorithm.
• Task 9 - Add spatial constraints to query builders for improved runtime performance.
• Task 10 - Finalize integration tests and prepare to push PR.
• Task 11 - Convert to WGS84 (for spatial binning and GeoServerIngestIT).
• MILESTONE - PR pushed, reviewed and merged.
• Task 12 - Handle tiling. This will be implemented after the initial PR is merged.

Tips for getting started:

• Task 1 and 2 - see Resources section below. And also, a close example would be: https://docs.geoserver.org/latest/en/user/styling/sld/extensions/rendering-transform.html#group-candidate-selection
• Task 3 - Similar to the .sld Subsample https://github.com/locationtech/geowave/blob/master/examples/data/slds/SubsamplePoints.sld#L15 as defined in GeoWave here: https://github.com/locationtech/geowave/blob/master/extensions/adapters/vector/src[…]ocationtech/geowave/adapter/vector/plugin/SubsampleProcess.java
• Task 3 - using hints: https://github.com/locationtech/geowave/blob/master/extensions%2Fadapters%2Fvector[…]ave%2Fadapter%2Fvector%2Fplugin%2FGeoWaveFeatureCollection.java

Resources:

• Task 1 - render transform
• Task 2 - heat map

Timeframe:

• 1 - 2 tasks per sprint or more as time allows.

[mzagorskirs]

Project ended early (this week); any remaining items for this epic will not be finished due to no more time/funding.