Kalshi US city temperature order books, live

Each row is the resting YES order book of one Kalshi temperature strike after each second, recorded live from the Kalshi websocket from 10 September 2026 for 24 US cities. One table per city series: daily high, daily low, and hourly temperature.

Coverage

• Window: from 2026-09-10 20:05 UTC onward, growing with every sealed segment, in UTC.
• Grain: one row per market ticker and second, holding the book after the last event of that second.
• Tables: 52, one per Kalshi series: 24 daily high-temperature series (KXHIGH…), 24 daily low-temperature series (KXLOW…) and 4 hourly temperature series (KXTEMP…HS), each named kalshi_high_<city>_orderbook, kalshi_low_<city>_orderbook or kalshi_temp_<city>1h_orderbook.
• Cities: Atlanta, Austin, Boston, Chicago, Dallas, Denver, Houston, Las Vegas, Los Angeles, Louisville, Miami, Minneapolis, New Orleans, New York City, Newark, Oklahoma City, Philadelphia, Phoenix, San Antonio, San Diego, San Francisco, Seattle, Trenton and Washington DC; hourly strikes for Chicago, Los Angeles, Miami and New York City.
• Cadence: each daily series opens about 12 strikes per day and each hourly series about 10 per hour; every table follows every open strike that carries a resting YES bid. Strikes near the forecast produce several rows a minute; far strikes produce a row only when their book changes.
• Freshness: each recording seals a segment every 300 seconds and each segment becomes a new version of its table.

Columns

• ticker — the Kalshi market ticker, for example KXHIGHNY-26SEP11-T86, naming the series, the day or hour, and the strike.
• timestamp_ms — the venue event time of the last book event in the second, in epoch milliseconds.
• sequence — the position of that event in the recording's arrival order.
• best_bid, best_ask — the best resting YES bid and ask, in cents from 0 to 100.
• bid_depth, ask_depth — the total resting contracts on the YES bid side and the YES ask side.
• yes_bids, yes_asks — JSON arrays of {price, size} for every resting level, best price first, in cents and whole contracts.
• ingest_time_ms — when the recorder received the frame, in epoch milliseconds.

Missing values

A side with no resting orders has a null best price, a depth of 0 and an empty ladder. A strike stops producing rows once it settles. Any minute in which a recording was not connected is a gap; nothing before a recording started can be recovered. International city series had no open markets on 10 September 2026 and have no table yet.

Suitable for

• Reconstructing the full YES limit order book of any US city temperature strike at any second.
• Measuring how the market prices a day's high or low as forecasts and observations arrive.
• Comparing the high, low and hourly books for the same city on the same day.
• Studying how order flow reacts to National Weather Service forecast updates and METAR observations.

Source and rights

Kalshi's public trade API websocket, orderbook_snapshot and orderbook_delta channels, recorded under the workspace's own Kalshi API key, one recording per series. Daily strikes settle on the National Weather Service daily climate report for each city's station; hourly strikes settle on the top-of-hour observation named in the market rules. Use is governed by Kalshi's API terms; no open licence is claimed by the publisher.