Project Lens is an alpha-level demo. The data sets are incomplete and imbalanced and algorithm parameters are being tuned. Results should thus be considered directional. Want to provide feedback, request features, or collaborate?

Project Lens is an exploration in viewing the world from different perspectives (or lenses). The goal in the small is to help people improve their epistimology and (meta-)rationality by understanding and diversifying their information sources. In the large, I hope to make a contribution to improving the information ecology, particularly in the realms of science, news, opinion, and public policy. We've experienced an explosion in information and communication in recent decades, but have not yet developed tools up to the task of consolidating the information into broadly accepted knowledge.

The first toy to play with is lenses based on Personalized PageRank over a Twitter endorsement graph (likes and retweets; despite people's claims about these not being endorsements, they're at least endorsements of interestingness). I have pre-configured many lenses for think tanks, journalistic institutions, field-specific Twitters (e.g. #ClimateTwitter, #EpiTwitter, #EconTwitter), and informal communities. These are implemented as personalization vectors for PageRank. Each lens has PageRank scores of Twitter accounts, which is interesting on its own, but these vectors can be used to weight URLs in tweets, which be used to generate aggregated, weighted scores for websites, blogs, books, podcasts, substack newsletters, and more.

I have many more experiments and toys on the drawing board. If anything you see here sparks ideas or would like to collaborate, please contact me.

Software Engineering


Informal Community