# DATASET: The whole SPX dealer-gamma archive, as files on your own machine.

> Every finished session minute by minute and strike by strike, in the volume, measured and open-interest books, as the same files the terminal reads. A new session lands the evening it closes. Every field is defined in the data dictionary, with its book, unit and the day it starts.

Source: https://firmtape.com/features/dataset

The DATASET licence, monthly or yearly. Reading single sessions on the site stays free.

Guide: [DATASET guide](https://firmtape.com/docs/dataset)

## The problem

**The free archive is built for reading one session at a time.**

- Research over a thousand sessions means a thousand page loads, and the site limits how many distinct days one visitor fetches.
- Three books sit side by side in every session, and a number from one is meaningless next to a number from another unless you know which is which.
- A backtest is only as good as the day it was last refreshed.

## With DATASET

**The archive as files, with a dictionary that names every field's book.**

The DATASET licence delivers every finished SPX session as one file a day, the same file the terminal reads, with the per-visitor limit on distinct days removed for your account. Each field's book is in its name and in the data dictionary, so the volume, measured and open-interest books are never mixed, and a session is added the evening it closes.

## Step by step

1. **Start with the dictionary**: Every field of the files and every LAB column, grouped by the book it reads. Read it before you license: it is public.

   ![Start with the dictionary](https://archive.firmtape.com/press/features/data/dictionary.eb9d819f.webp)
2. **Pick a book**: VOLUME, MEASURED, OPEN INTEREST or NO BOOK: each tab lists the fields of one book, and the volume book is the default the chart draws.

   ![Pick a book](https://archive.firmtape.com/press/features/data/dictionary.eb9d819f.webp)
3. **Read a field**: Each row gives the field, what it means, its unit and since when it exists. Its suffix names its book.

   ![Read a field](https://archive.firmtape.com/press/features/data/dictionary.eb9d819f.webp)
4. **Download every evening**: With the licence active, fetch each day's file with your account's token; a new session lands the evening it closes.

   ![Download every evening](https://archive.firmtape.com/press/features/data/dictionary.eb9d819f.webp)

## What you can do with DATASET

- **Every session as a file**: One file a day, the same one the terminal reads, minute by minute and strike by strike.
- **Three books, never mixed**: Volume, measured and open interest, each field named for its book.
- **A new session every evening**: The archive grows by one session each evening, so a study rerun tomorrow includes today.

## Everything in DATASET

6 features in 2 groups.

### The files

- [The day files and the dictionary](https://firmtape.com/docs/dataset#archive-files): Compute on every finished session by fetching one file a day, each field defined in the data dictionary.
- [Three books in every file](https://firmtape.com/docs/dataset#three-books): Keep the volume, measured and open-interest books apart by reading the book each field names.
- [A new session every evening](https://firmtape.com/docs/dataset#evening-delivery): Keep a study current by fetching the session that lands each evening after the close.
- [Older days in the levels API](https://firmtape.com/docs/dataset#deep-history): Ask the levels API for days older than its usual window by holding the licence on the same account.

### The licence

- [Dataset licence](https://firmtape.com/docs/dataset#dataset-licence): Run your own research on every finished session by licensing the archive as files.
- [Licensing for your product](https://firmtape.com/docs/dataset#licensing): Show FirmTape's levels or stream to your own users by taking a printed redistribution licence.

## What it gives you

- **Your own compute**: No credits, no browser and no rate limit on distinct days: the files run in your own notebook or pipeline.
- **Three books, never mixed**: Every field names its book by suffix and in the dictionary, with its unit and the date it starts.
- **Current every evening**: The session lands the evening it closes, so a study rerun tomorrow includes today.

## Add to DATASET

- [Redistribution licences](https://firmtape.com/licensing): Show FirmTape's levels or its stream to other people in your app, site, bot or channel, on printed terms

## Works with

- [REPLAY: The session archive and session pages](https://firmtape.com/features/replay/session-archive): Read any session on the site for free before you license the files.
- [LAB: The LAB vocabulary](https://firmtape.com/docs/lab#vocabulary): LAB's columns are built from the same fields, so a rule tested in LAB can be rebuilt from the files.
- [LIVE: Levels API](https://firmtape.com/features/live/levels-api): The levels API answers recent history for charts; older days come with the licence.

## Questions about DATASET

**What is in the files?**

Every finished SPX session minute by minute with per-strike detail in the volume, measured and open-interest books, plus the market calendar file.

**How do I download?**

One file a day at the same address the terminal reads, with your account's token as a bearer header; the day list tells you which days exist.

**How fresh is it?**

A session lands the evening it closes; delivery starts the evening the licence starts.

**Can I show the data to my own users?**

Not under this licence, which is for your own research. Showing the numbers to other people is a redistribution licence, priced on /licensing.

**Does it include the raw options tape?**

No. Only FirmTape's derived measurements; licensed raw data never leaves FirmTape.

**Does it carry today's session as it trades?**

No. The files are finished sessions; today's session as it trades is LIVE.

### What it does not do

- It does not include the raw options prints or quotes.
- It does not cover today's session while it trades.
- It does not allow redistribution or a competing service; that needs a separate licence.
- It does not forecast anything: every field is a measurement of a finished session.
