Eduardo Furtado

Graph Analysis in App Navigation

Featured image
Image generated with NetworkX and Pyvis

I worked on a project to better understand how users interact within a super app, which combines services like parking payments, online shopping, buying movie tickets, loyalty programs, to name a few. Despite offering many features, users often stick to just using one and not exploring more of what was available.

To find out which sequences of app features users commonly follow and identify underused paths, I used graph analysis to build a directed graph using Python with NetworkX, where each node represents a feature within the app, and the edges show how users move from one feature to another within a certain time frame. The size of each node reflects the number of unique users for that feature, and the thickness of the edges indicates the proportion of users transitioning between them.

An interactive visualization was created with Pyvis, which allowed users to explore the data by hovering over nodes and edges to see detailed results and get insights for improvements in app design and communication strategies.