A recipe for pulling real-time crypto market data from Binance's public APIs. No account, no API key, no authentication — every call is a plain HTTPS GET. The only restriction is geography: Binance blocks Australia and the US (HTTP 451). Singapore and Hong Kong are confirmed working.
Two data sources answer two different questions:
Spot cohort analysis answers: who is buying and selling, by size? Are whales accumulating or distributing? Is retail leading or following? It works by walking every trade on the spot market, sorting each by dollar value into size buckets (retail under $1k, small $1k–$10k, mid $10k–$50k, large $50k–$250k, whale over $250k), and totalling buy vs sell dollars in each bucket.
Derivatives positioning answers: how are futures traders positioned? Are shorts or longs crowded? Is money entering or leaving? What does it cost to hold? It uses eight single-call endpoints on the futures API covering open interest, funding rates, long/short ratios by account count and by dollar weight, and taker aggression.
The one rule learned the hard way: never combine large and whale into "big money" — they regularly move in opposite directions. A 14-day BTC dataset showed large was -$116M while whales were +$45.6M. Always report each tier separately.
How to read the output together: compare spot buyers and derivatives buyers. Are they pointing the same direction? Is OI building while spot is accumulating (conviction) or is OI building while spot is distributing (divergence)? That divergence is where the signal lives.
Rate limits: 1,200 requests/min on spot, 2,400/min on futures. The spot cohort walk is slow because of trade volume, not rate limits — BTC can have 500k+ trades per day.
Symbols: append USDT to the coin name. Spot: ZECUSDT, BTCUSDT, LTCUSDT, XRPUSDT. Futures: same, except PEPE is 1000PEPEUSDT.