US Equities Trade and Quote
One row is one consolidated-feed (SIP) event for a US-listed equity: a venue or NBBO bid/ask quote update, a trade, or a trade cancellation, stamped to the nanosecond with venue and condition flags. Used to replay the order-of-arrival market state for microstructure research and backtests.
How to use it
Times are local US Eastern clock times (DST-aware), not UTC. Treat TRADE NB / NBBO-flagged rows as the backtest-grade subset. Cancellations reference earlier trades and can arrive hours later, so aggregate only after the session closes. Prices are unadjusted unless the dataset name says Adjusted; join us-equities-*-adjustment-factors on SecId for adjustment.
Typical use
Backtesting, execution and microstructure research, feature engineering for US equities.
Limitations
Delivered as recorded from the source feed: AlgoSeek does not remove crossed or locked quotes, bad prints, or out-of-sequence packets unless contracted to. Consumers own cleaning. Sample channel: S3 CSV (sample_channel s3/cloud_storage, eq_taq.csv, 100 rows, AAPL on 2023-08-02); this dataset has no ArdaDB table sample. The JSON preview is identical to the CSV. Closed EventType set per guide Table 3: QUOTE BID, QUOTE BID NB, QUOTE ASK, QUOTE ASK NB, TRADE, TRADE NB, TRADE CANCELLED, TRADE NB CANCELLED; the 100-row sample contains only the first six. Do not sum quote sizes across NB and non-NB rows: an NB row repeats a quote already published by the originating exchange. TRADE NB is a trade at or within the NBBO; the guide recommends TRADE NB for backtesting. Trades away from the NBBO include some FINRA TRF reports, late reports, extended-hours trades (no NBBO pre/post), ISO orders and other trade-through exemptions.
Coverage
All trades and level-1 bid/ask quotes from the SIP consolidated feed (CTA Tape A/B and UTP Tape C) for listed stocks, ETNs, ETFs, ADRs and funds, from 15+ US exchanges and marketplaces (Table 6 lists 21 venues incl. FINRA, plus UNKNOWN/INVALID). History start is not stated; nanosecond timestamps start in 2016, millisecond before.
Venues (23)
- NYSE
- ARCA
- AMEX
- NSE
- CSE
- Nasdaq
- Nasdaq BX
- NASDAQ PSX
- ISE
- CBOE
- BATS Y
- BATS
- EDGA
- EDGX
- IEX
- LTSE
- MIAX
- MEMX
- FINRA
- NE24X
- TXSE
- UNKNOWN
- INVALID
Answers to common questions
- What time zone are the timestamps in?
- US Eastern local time with daylight saving (the guides label it EST/ET).
- Are prices adjusted for splits and dividends?
- Only in datasets whose name contains Adjusted; all others are as traded.
Usage constraints
- Licensed vendor data: never export raw rows to a tenant-local destination or include them verbatim in a public artifact; serve only through entitlement-checked platform reads.
- Do not treat the vendor's `Ticker` as a stable identity across time; join on an AlgoSeek security key (SecId or ASID, which are distinct) or the platform symbol mapping.
Known issues (9)
- Primary key is inferred; AlgoSeek's guide does not name one.
- Documentation discrepancyConditions width/case: guide Appendix A says flags are eight-character hex and Tables 1/4/5 print them zero-padded (00000001, 80000401); the CSV sample prints lower-case hex with leading zeros dropped (value lengths 1, 3, 7, 8 such as '1', '401', '4000001', 'a0000021').
- Documentation discrepancyExchange spelling: guide Table 6 values 'Nasdaq', 'Nasdaq BX', 'MIAX' appear in the sample as 'NASDAQ', 'NASDAQ BX', 'MIAX PEARL'; the guide itself mixes case ('Nasdaq' vs 'NASDAQ PSX').
- Documentation discrepancyGuide text before Table 6 says '19 SIP Market Center IDs (including FINRA)' but the table lists 21 venues plus UNKNOWN and INVALID; the intro says '15+ exchanges'.
- Documentation discrepancyTime zone is labelled 'EST' in the guide although the 04:00-20:00 session and 09:30/16:00 market hours are local Eastern clock times; DST handling is not stated.
- Documentation discrepancyGuide Table 2 has no 'Missing' column (other algoseek guides do), so null/empty behaviour per field is undocumented; the 100-row sample has no empty cells.
- Documentation discrepancyNBBO paragraph describes a 'crossed NBBO (e.g. the Bid = Offer)'; Bid = Offer is a locked market, crossed means bid > offer (the Daily TAQ guide defines both correctly).
- Documentation discrepancyTrade Only guide Table 3 defines TRADE NB as a trade 'at national best price' and omits TRADE NB CANCELLED; this guide defines TRADE NB as at or within the NBBO and lists TRADE NB CANCELLED.
- Documentation discrepancyNo history start date is given in the guide or on the product page; the sample covers one session slice (AAPL, 2023-08-02, from 09:30:00.000815872).
8 columns in delivery order · file or table name eq_taq. Types are ODCS logical types; the format line shows how values are written.
| Column | Type | Key | Description |
|---|---|---|---|
| Date | dateyyyymmdd | Key 1 · Partition · Required | Trading date of the event, written yyyymmdd. Source: vendor guide and sample |
| Timestamp | timeHH:MM:SS.mmmuuunnn (ns) from 2016; HH:MM:SS.mmm (ms) before 2016 | Key 3 · Required | Event time of day in US Eastern: nanosecond precision from 2016, millisecond before. Import as text, not an Excel time. Source: vendor guide and sample |
| EventType | string · enum (8) | Key 4 · Required | Kind of event: venue or NBBO-creating bid/ask quote, trade, or a trade cancellation. Source: vendor guide and sample 8 allowed values
|
| Ticker | string | Key 2 · Partition · Required | Symbol name of the security. Source: vendor guide and sample |
| Price | number | Price of the bid, ask or trade; up to 4 decimals for sub-penny prices; 0.00 on zero-quote closing events. Unit: USD · Source: vendor guide and sample | |
| Quantity | integeri64 | Number of shares in the quote or trade; can be 0 for some event types (e.g. zero closing quotes). Unit: shares · Source: vendor guide and sample | |
| Exchange | string · enum (23) | Key 5 · Required | Exchange or reporting venue that produced the event, as an algoseek venue code. Source: vendor guide and sample 23 allowed values
|
| Conditions | stringhex uint32 bitmask, lower-case, no 0x prefix; guide examples zero-padded to 8 chars, sample drops leading zeros | Bitmask of sale-condition flags (trades) or quote-condition flags (quotes), as base-16 text. Source: vendor guide and sample |
- errorrow_count_positivecompleteness
Every delivered partition (trade date, or snapshot) carries at least one row.
rowCount must be greater than 0
- warningprimary_key_uniqueuniqueness
Best-effort key for an event stream: same-timestamp events from one venue can repeat, so duplicates are reported, not rejected.
duplicateValues (properties: Date, Ticker, Timestamp, EventType, Exchange) must be 0
- errorQuantity_non_negativeaccuracy
Quantity is never negative.
SELECT COUNT(*) FROM {object} WHERE Quantity < 0 - errorconditions_hex_uint32conformity
Conditions is a hex uint32 bit mask without a 0x prefix. The guide shows 8 zero-padded characters; delivered files use lower case and drop leading zeros (e.g. 401, a0000021).
invalidValues (pattern: ^[0-9A-Fa-f]{1,8}$) must be 0
Delivery
S3 flat files
AlgoSeek flat-file delivery (the layout the schema below describes).
- Layout:
- Plain-text CSV with a fixed header row; by default one file per symbol per trading day (all events for e.g. AAPL on one date in one file); gzip-compressed (.csv.gz, ~8:1). The guide gives no bucket/key path for the data files; only the calendar files s3://us-equity-market-holidays/holidays.csv and earlycloses.csv are named.
- Compression:
- gzip
- Access:
- Bucket access is granted per subscription; bucket names arrive with license provisioning.
REST API
AlgoSeek Datasets REST API (X-API-KEY header; JSON, CSV or gzipped CSV). Returns AlgoSeek's normalized API layout, which differs from the flat-file columns in this contract's schema.
https://api.algoseek.com/v1/data/us-equity/taq/:trade_date/:identifier
- Authentication:
- X-API-KEY header issued by AlgoSeek
- Response formats:
- json, csv, csv_gzip
- Pagination:
- offset + limit; tick endpoints take trade_date and identifier (Ticker or ASID) in the path
Service levels
- Update frequency
- Daily
- Retention
- Bounded by the license term
- Retention of landed copies is bounded by the AlgoSeek license; purge on termination.
- History starts
- Not stated in the vendor documentation
Service levels describe the vendor feed as documented. They are draft values, not a commitment by AlphaSwarm.
Data Contract Editor · read-only preview — edits are not saved
Open-source Data Contract Editor (MIT). Changes stay in this browser tab: Cancel restores the published contract, and Download copy saves your edited YAML to your device only.
Contracts follow the Open Data Contract Standard (Bitol, LF AI & Data).
Vendor names are trademarks of their owners.