OpenRank: Powering Apps with Contextual and Personalized Graph Feeds

The article is a tutorial by an investor and advisor to the OpenRank team that presents OpenRank's eigentrust algorithm as a way to build personalized feeds for crypto apps. It says the algorithm ranks nodes in a graph by distributing trust, using a pretrust seed set, a localtrust graph of interactions, and a weighting constant c. In the worked example, the author builds a feed of recommended contracts from the wallets of Farcaster casters in the /base channel, using the top 100 casters by engagement as pretrust nodes. The graph covered 65,755 users, and the pretrust nodes reached 19% of it. The author also announces a Bytexplorers research mission in July to build graph datasets for feeds.