Private Polymarket Scraping Competency Task
Budget: $10 – $30 USD
I’m opening this room so we can collaborate one-on-one on the assignment we previously discussed.
Here is a mini-project that is a small test before committing to a larger project.
Part A — Sports Market Identification (uses my existing files):
Parse the full gamma_ALL_events.jsonl file (264K events, 2.8 GB) and extract every unique tag across all events. Deliver the complete tag list and identify which ones are sports. Then produce the definitive list of sports markets for a single sport only — pick NBA. Give me every NBA market with its conditionId, clobTokenIds, event title, market question, outcomes, and settlement result. This proves you can work with the existing data, understand the event-market-tag hierarchy, and handle the stringified JSON quirk.
Part B — On-Chain Trade Extraction (small sample):
Pick one specific NBA event from Part A that has settled and has a reasonable number of trades. Extract every OrderFilled event for that event's token IDs via eth_getLogs on a Polygon archive node. Decode each fill and deliver a structured file with maker, taker, token ID, amounts, price per share, timestamp, and tx hash. Match each fill to the correct market and outcome side. This proves you can query the chain, decode the events correctly, filter by token ID, and structure the output.
What I'm checking: That the tag extraction is truly complete and not sampled, that the NBA market list is comprehensive, that the on-chain fills are correctly decoded with prices that make sense (between 0 and 1), that maker/taker addresses are real wallets and not the contract address, and that the output schema matches what we agreed on for the full project.
Here is a mini-project that is a small test before committing to a larger project.
Part A — Sports Market Identification (uses my existing files):
Parse the full gamma_ALL_events.jsonl file (264K events, 2.8 GB) and extract every unique tag across all events. Deliver the complete tag list and identify which ones are sports. Then produce the definitive list of sports markets for a single sport only — pick NBA. Give me every NBA market with its conditionId, clobTokenIds, event title, market question, outcomes, and settlement result. This proves you can work with the existing data, understand the event-market-tag hierarchy, and handle the stringified JSON quirk.
Part B — On-Chain Trade Extraction (small sample):
Pick one specific NBA event from Part A that has settled and has a reasonable number of trades. Extract every OrderFilled event for that event's token IDs via eth_getLogs on a Polygon archive node. Decode each fill and deliver a structured file with maker, taker, token ID, amounts, price per share, timestamp, and tx hash. Match each fill to the correct market and outcome side. This proves you can query the chain, decode the events correctly, filter by token ID, and structure the output.
What I'm checking: That the tag extraction is truly complete and not sampled, that the NBA market list is comprehensive, that the on-chain fills are correctly decoded with prices that make sense (between 0 and 1), that maker/taker addresses are real wallets and not the contract address, and that the output schema matches what we agreed on for the full project.