I’ve been maintaining a comprehensive Bluesky dataset in BigQuery and am looking to license access to cover infrastructure costs on a hobby basis. Due to the nature of Bluesky and the underlying ATProto, this includes all posts, follows, likes, etc.<\/p>\n
Unfortunately, it’s gotten expensive, and I’m going to have to shut it down if I can’t find a way to reduce the cost. <\/p>\n
What’s available:<\/strong> – ~11.4 billion raw events – Full historical coverage from Bluesky’s launch, backfilled from ATProto CAR file repositories and normalized into a single unified schema – Ongoing live stream via Jetstream – Raw CAR backfill table also available separately if useful – BigQuery-native access \u2014 no ETL on your end<\/p>\n Unpacked tables include:<\/strong> – Posts (with hashtags, links, mentions) – Likes, reposts, follows, blocks – Deletes – Profile updates – Follower\/friend graph materialized views<\/p>\n Who this might be useful for:<\/strong> – Researchers studying decentralized social networks, post-Twitter migration, or online discourse – Media intelligence \/ social listening products – ATProto developers who want query access to the full event history<\/p>\n Since this is in BigQuery, you can do joins, which leads to all kinds of fun queries like “Give me all the accounts most overfollowed by the unique followers reached by posts mentioning “Chartreuse Goose” for all time.” A query like that would run in 15-30sec.<\/p>\n Also 100% open to releasing to the community if we can find a way to pay for it.<\/p>\n Anyone interested? Not trying to turn a profit here — just trying to keep a resource online. (Hope that’s OK for the rules here!)<\/p>\n<\/div>\n submitted by \/u\/aboothe726 <\/a>
[link]<\/a><\/span> [comments]<\/a><\/span><\/p>
