Graphing Call Distributions by Country using 3D.js
When dealing with VoIP interconnections, you need to analyze the inbound and outbound call traffic from and to various countries for constantly negotiating termination fees you might receive from your peering partners, and origination fees you have to pay for.
Having an overview of your most common origination and termination countries significantly helps with figuring out where to focus on, so we started to experiment with different options how to display this information in a compact and useful way.
During our evaluation of different graph types, we came across Chord Diagrams and an example for displaying Uber Rides by Neighborhood.
Using a chord diagram, you can encode the relative frequency of a connection between two points, which in our example would be calls from one country to another and vice versa. Also, the connections are directed, so you can see immediately which of the two end points (countries) has more originations. So applying the example from above to CDRs generated by the Sipwise NGCP and integrating it into the NGCP Panel, we end up with a graph like this:
You can hover over the different country codes to highlight a certain connection, and hovering over the connections itself shows the absolute number of calls for each direction between two specific countries.
The new Call Distribution statistics will be available in the upcoming v3.2 version of the NGCP.