Customize Kibana Region Maps

With GeoJSON Boundaries

(λx.x)eranga

Published in

Effectz.AI

5 min readJun 1, 2021

Background

On my previous post I have discussed about using different visualizations(graphs, charts, matrixes, tables etc) with Kibana. In this post I’m gonna discuss about visualizing the data with Kibana Region Maps. Region Maps in Kibana are a better solution to represent a whole area or region. All the source codes which related to this post available in gitlab. Please clone the repo and continue the post.

Run Elasticsearch

To run Elasticsearch I have used Elassandra docker image. Elassandra built by combining Elasticseach with Cassandra. It comes Elasticsearch as a Cassandra plugin. Basically it has Cassandra API as well as Elasticsearch API. When data save on Cassandra it will automatically index on Elasticsearch. Read more about elassandra from here. Following is the docker-compose.yml to run the Elassandra.

Elasticsearch Index

I have a Elasticsearch index vote. It contains mock election result data of Sri Lankan elections in past few years. To create the index and migrate sample data into the index, I have used a separate service Aplos. When running Aplos service, it creates Elasticsearch index and populate data on to it. Following is the docker-compose.yml to run the Aplos service.

Following is the structure of the vote Elasticsearch index. This index can be configured in Kibana. There is a field district in the vote index which I have used to build region maps.

GeoJSON Geo-Boundaries

Kibana provides region maps of several default regions(e.g World Countries, US States, Australia States etc). To have maps with custom regions(regions which does not provide in default Kibana) we need to get the geo-boundaries of the regions as GeoJSON format. GeoJSON is an open standard format designed for representing simple geographical features, along with their non-spatial attributes, based on the JSON format. Once get the GeoJSON boundaries of the regions, we can overlay them on the Elastic Map Service to display aggregations for custom geographical areas natively in Kibana. In this example I’m using Kibana Maps to visualize election results of different districts in Sri Lanka. For that, first I have obtained the geo-boundaries of Sri Lankan districts into GeoJSON file named electorate.json.

GeoJSON Http Server

To build custom maps in Kibana, we need to host our GeoJSON files in CORS enabled HTTP/HTTPS web server. So in here I have built simple CORS enabled web server with nodejs http-server package to serves the electorate.json file. Following is the Dockerfile and docker-entry.sh I have used dockerize http-server.

Following is the docker-compose.yml entry to run the dockerized nodejs http-server instance.

Kibana Config File

Next I have defined the information in the electorate.json GeoJSON file(url, fields etc) in the Kibana config. Following is the kibana.yml config file which defined the electorate.json GeoJSON file information.

The URL of the hosted GeoJSON file is http://geo:800/electorate.json. In here the geo is the host name of the http-server service I have built to host the GeoJSON file. The GeoJSON file contains a filed called electoralDistrict which defines the district name. In the Kibana config file I have defined that field information as well. Following is the docker-compose.yml entry to run the Kibana with given config file. In here I have defined a link to http-server(geo). It will add geo host name entry into the /etc/hosts file of the Kibana docker container. Then Kibana service can serve the http-server via the geo host name.

Run services

Now I can start the services in the following order. The Aplos service will create Elasticsearch index and bootstrap data into it.

Configure Kibana Index

To create the map, first I need to define the Elastisearch index in the Kibana, it can be done in following way.

Create Map

Then I can create a region map with vote index by joining the data on the electorate.json GeoJSON file. When creating the region map, I need to specify the field(identified as Join field) in the Elasticsearch index(district) which matches to the field in the GeoJSON file(electorateDistrict). This field matching can be define from the Buckets shape field aggregation in the region map configuration. Following are the main configuration I have done in the map.

Aggregation Matrix Sum of party_votes
Bucket Shape field Terms district
Layer settings Vectors map Sri Lanka Districts
Join field district

Once created the map with these configurations, Kibana will visualize the sum of party_votes in different districts in Sri Lanka with different colors. Kibana provides different color themes(e.g greens, reds, yellow to read, green to red etc) to visualize the data.