Calcifer Computing letterhead Calcifer Computing letterhead

Whatever ultramodernity places under the dominion of signs postmodernity subverts with virus… no longer what does it mean? but how does it spread?

What is a meme?

The datasets

DatasetScaleWhat it isDownload
US Congress tweets 5.1M tweets 5,095,245 tweets from members of the US Congress and executive branch, 1999–2026. Each row: tweet id, author handle and name, party, chamber or role, state, timestamp, full tweet text, and a topic label where available. aws s3 cp --no-sign-request s3://calcifer-hot/hopkins-hackathon-2026/congress-tweets-unified.parquet .
Twitter firehose 395M tweets 395M tweets from across X (all languages) posted in the trailing month. Each row: tweet id, author id, text, timestamp, language, reply / quote / conversation ids, and like, reply, retweet, quote, view and bookmark counts — with repeat snapshots per tweet, so engagement is a time series. aws s3 cp --recursive --no-sign-request s3://calcifer-hot/hopkins-hackathon-2026/twitter-firehose-last-month/ ./twitter-firehose/
Parkour Instagram Niche 451k posts
10.8M observations
451,301 Instagram posts from ~14,900 accounts in the parkour / freerunning niche: captions, media URLs, hashtags, and like / comment / view / share counts, with 10.8M timestamped engagement snapshots and a 595,730-edge account-to-account graph. aws s3 cp --recursive --no-sign-request s3://calcifer-hot/hopkins-hackathon-2026/parkour-instagram-niche/ ./parkour/
TikTok videos 4.5B videos 4.5 billion TikTok videos, one row each. Per video: caption, the sound used (music_id, a join key for tracking a sound across videos), play / like / comment / share / save counts, duration, country, language and post time. Sharded into 27 Parquet files of ~10 GiB — the split is a hash of the creator ID, so any single shard is an unbiased sample of roughly 167 million videos. Start with one. One shard · ~10 GiB · start hereaws s3 cp --no-sign-request s3://calcifer-hot/tiktok/tiktok-videos-4b/videos-00.parquet .All 27 shards · 270 GiBaws s3 cp --recursive --no-sign-request s3://calcifer-hot/tiktok/tiktok-videos-4b/ ./tiktok/

Judging Criteria

Vibes

Judging is purely vibes based. $1,000 in cash to the team that articulates the most interesting memetics phenomenon. Use these datasets or bring your own; anything you can pitch as memetics is fair game.


Happy Hacking!

Interested in learning more? Sign up for Introduction to Computational Semiotics during intersession!