ByteCycler is a crowdsourced sensemaking platform built by Johann Cooper and I. It started off as a hackathon project, and has since then turned into a full-fledged web platform. The platform generated graphs of the public discourse around a current issue, showing resonant and dissonant perspectives (nodes), as well as the relations (edges) between them.
I worked on 4-5 prototypes of the platform over the course of two years while attending full time studies, and got acquainted with full-stack development in the process (React/Node/Express/SQL, & Django).
At its core is the Echo Score and the Topic Graph (where each article is represented with a node).
The Echo Score of an article shows how much the snippets in an article are echoed by other articles. It's the percentage of supported snippets out of the article's total pool of supported and contradicted snippets. Nodes and connections are color coded from blue (highest score) to yellow (lowest score).
Here for example, the maroon article has a lower echo score than the blue ones. Connected articles disagree with the maroon one more.
If we hover over that node, we can see that most of that disagreement comes from the top node, since the echo score between those two nodes is yellow - almost 0.
We also had pages for every supported source / news outlet. Users could see all articles linked for that source, and rate that source’s credibility.