Mining the Devoxx Network

Thursday, October 24, 2019

JUVO specialises in data productivity improvement projects, with a focus on data intelligence and software engineering. We asked ourselves, how can we connect our two specialisations?

Devoxx has many different talks. Each of those sessions discusses technical and less technical subjects. But how do all these topics relate to each other? To discover that, we built a network discovery algorithm and executed this on the Devoxx program website. Below you can see the result showing how the top 20 found topics connect to each other. Each circle shows a connection (i.e. they are tackled in the same session) between the linked topics. The bigger the circle, the more this combination occurs across sessions. 

We also crafted an interactive version, go to http://rpubs.com/meskensj/devoxx-demistified for more! 

Interesting to see is how kubernetes and serverless are the most interwoven topics of Devoxx, while UI seems to be the linking pin between many different types of technologies

Version 2 (where you see Java a bit better):

Toolkit

This network was built in R using the RScraper and Arules packages. The algorithm beyond those packages is apriori to find association rules across talks. Colours in the network above indicate the “lift” of the associations.

We also crafted an interactive version, go to http://rpubs.com/meskensj/devoxx-demistified for more!