
Track GIS Technician Work with the ELK Stack – Part 3
Part 3: Use Kibana to Visualize Data from Elasticsearch
In this final part of the series, I’ll show you visualizations that can be done with Kibana using data in Elasticsearch. This is the fun part and there’s no coding required!
Note: There are countless visualizations you can make in Kibana, but these are a few of our favorites for the purpose of monitoring a digitizing project
One interesting and useful visualization for looking at digitized objects is the Coordinate Map. It aggregates our location data from Elasticsearch into a map of your choice. Below we’ve shown a graduated symbol map, where the size and color of the circles refer to the density of objects marked in a region.
This could be helpful, for example, if you intend to detect forest cover, but your training data is located over deserts, whcihich would not yield accurate results at all. In this case, you would know right away that the dataset would not fit your needs from the visualization (which is updating every minute, or however long you’ve specified in Logstash) and notify your digitizers to divert their efforts appropriately, saving time and resources. On the other hand, if you want your object markings to be dispersed heterogeneously around the globe, the visualization would be able to confirm as well.
Here’s another scenario. We are curious about how productive our digitizers are being on an hourly basis. Kibana is able to aggregate our Elasticsearch data by the timestamp attribute we populated our features with. This gives us an hourly time series graph with the number of objects digitized every hour of the workday since our project began a few weeks ago:
From the visualization above, we can see where the peaks and valleys of the work day are and how many objects were counted for each hour of each day of the project. We can see which days and hours are more productive than others.
We can aggregate our Elasticsearch data by count and object type as well. This bar graph shows the total counts of each of the objects:
We can also aggregate by count and person. This one shows each bar as an editor and the number of objects counted by that editor. Note: names changed to protect identity of our digitizers
We can even create a split series, where we get a breakdown of which objects comprise each users’ total count:
In the line graph below, we’ve created a split series where each line is a digitizer. The x-axis referes to the date and the y-axis refers to the counted objects. These are the cumulative object counts since the beginning on the project. We can see which users have been collecting the most objects, and which users are working at the fastest rates.
We can use the metric and goal visualizations to display important numbers too. We can display today’s total number of digitized objects as well as the total number of digitized objects since the beginning of the project.

15,568 features digitized today using the metric visualization (left), 433,330 objects digitized in total using the goal visualization (right). The goal visualization also shows the fraction done in comparison to a goal number of digitized objects by the end of the project
Here are all the examples we just discussed, together in a dashboard:

The above dashboard shows project metrics with visualizations created using Kibana. Names redacted/removed to preserve anonymity.
Summary
In this last section, we talked about a few visualizations that are possible using Kibana. A dashboard like this one will allow you to monitor all kinds of metrics easily and quickly, such as digitizers’ progress, hourly rate, total counts, object type counts, and geospatial distribution.
No Comments