Kaggle NYSE S&P 500 Historical Prices & Fundamentals

Two things every equity model wants - where the price went and what the filings said - in one package. The Kaggle NYSE S&P 500 Historical Prices & Fundamentals dataset bundles four CSVs into a single archive: two parallel daily price tables, an annual fundamentals table built from SEC 10-K filings, and a company reference sheet with GICS classifications.

The numbers that earn it shelf space: roughly 851,000 price rows across about 500 tickers, spanning January 2010 through December 2016; a second copy of that same history adjusted for the roughly 140 splits the window contains; about 1,400 annual fundamental rows carrying approximately 78 line items per company-year for fiscal years 2012-2016; and a reference table of roughly 500 companies with sector, sub-industry and CIK.

Within Datadory's catalog of 1,744 datasets across 159 viable industries, this is the only S&P-500-specific record that puts prices and filing-derived fundamentals side by side on the investment banking & brokerage shelf - the rest of that shelf splits into regulator feeds and statistics series.

Get a sample of this dataset

Captured during the August 2026 research pass, straight from the tables:

# daily price row (raw series)
date       2016-01-05
symbol     WLTW
open       123.43
close      125.84
low        122.31
high       126.25
volume     2163600

# company reference row (securities)
ticker_symbol          MMM
security               3M Company
gics_sector            Industrials
gics_sub_industry      Industrial Conglomerates
address_of_headquarters St. Paul, Minnesota
cik                    0000066740

# annual fundamentals row (truncated at the source mirror)
ticker_symbol     AAL
period_ending     2012-12-31
accounts_payable  3068000000
net_income        (null in mirror)

Reading the rows: the WLTW bar shows why recent listings matter - Willis Towers Watson only enters the tape near the end of the window, so its history starts in 2016 rather than 2010. The MMM reference row carries the classification spine; its CIK is the bridge that lets a fundamentals figure trace back to the exact 10-K it came from. The AAL row is shown truncated because the source mirror cuts off mid-record - your sample arrives complete, with nulls flagged rather than silently zeroed.

Eight field groups verified during the August 2026 review. The spine first - the long tail of roughly 78 annual line items sits folded under additional fields on request.

• date (date) - Trading day of the price observation in the raw and split-adjusted price tables. Example: 2016-01-05.
• symbol (string) - Ticker symbol of the listed company; the join key between price history, fundamentals and the reference table. Example: WLTW.
• open / close / low / high (number) - Daily opening, closing, intraday low and intraday high prices in dollars. Example: 123.43.
• volume (integer) - Shares traded during the session. Example: 2163600.
• Ticker Symbol / Period Ending (string / date) - Composite key in the fundamentals table identifying the company and the fiscal period end of its 10-K metrics. Example: AAL, 2012-12-31.
• GICS Sector / GICS Sub Industry (enum) - Sector and sub-industry classification carried on the company reference table. Example: Industrials / Industrial Conglomerates.
• CIK (string) - SEC Central Index Key, enabling joins from a fundamentals figure back to its EDGAR filing. Example: 0000066740.
• Accounts Payable ... Total Revenue (~78 columns) (number) - Annual 10-K financial-statement line items: accounts payable and receivable, capital expenditures, cash ratios, cost of revenue, EBIT and pre-tax earnings, goodwill, gross margin and profit, income tax, intangibles, interest expense, inventory, liabilities, long-term debt, net cash flows by activity, net income lines, operating income, profit margins, R&D, retained earnings, total equity, total revenue and valuation ratios. Example: Net Income, Gross Margin, EBIT.
• Additional fields - per-column coverage counts, any extended identifiers beyond the verified spine, and column availability confirmed per fiscal year before modeling. Confirmed with your sample rather than promised blind.

The dictionary below is the same spine in table form. Ask for folded fields with your sample and they arrive confirmed in the same tables, not a separate negotiation.

Three chips, honestly drawn:

• Geography - United States - S&P 500 constituents listed on the NYSE
• Temporal - Prices January 2010 - December 2016 (shorter histories for recent listings); fundamentals fiscal years 2012-2016
• Granularity - One price row per ticker per trading day; one fundamental row per company per fiscal year

One scope rule worth knowing before you model: this is a fixed historical window, not a rolling one. Every date range you build on it ends in December 2016, so pair it with a live feed wherever recency matters - and enjoy perfectly reproducible backtests everywhere else.

API, files, or your warehouse. Daily, weekly, or hourly.

The four-table structure lands however your stack wants it - pushed to storage, served over an endpoint, or synced straight into your database with the price and fundamentals tables already keyed for the join. Name the tickers, the field set and the history window when you request the sample; the sample ships first either way.

One honest note on cadence: this corpus is a fixed historical snapshot, so scheduled deliveries matter for keeping derived features reproducible across your systems - not for catching new bars. Nothing here updates past December 2016, and pretending otherwise is how silent train/serve drift starts.

A frozen seven-year panel with filings attached earns its keep in specific jobs:

• Data scientists & ML engineers - start regression, factor and ML tutorials on split-adjusted daily bars joined to 10-K fundamentals, with every run reproducing bit-for-bit (data scientists use cases).
• Developers & data-product builders - demo portfolio, charting and valuation products against real tickers without stitching a price feed together first (developers & builders use cases).
• Journalists, academics & students - teach data cleaning, joins and valuation concepts on a widely forked corpus whose quirks are documented rather than discovered (journalists & academics use cases).
• Competitive intel & product teams - prototype screening logic on real symbols before committing to a licensed feed (competitive intel use cases).
• Investors & quant researchers - keep it as the fixed control panel for benchmarking live strategies against an era everyone can see (investors & quants use cases).
• Market researchers & consultants - illustrate valuation-ratio mechanics with real numbers in client explainers, flagging the 2016 cutoff up front (market researchers use cases).

Three things upfront. First, the split story: the raw price table deliberately ignores the roughly 140 splits in the window, so any return math runs against the split-adjusted table - using the raw one produces fake crashes on every split date. Second, survivorship: prices were sourced from Yahoo Finance without survivorship handling, meaning today's constituents dominate the tape and companies delisted along the way may not be there - reconstructing a point-in-time index membership needs care. Third, scope of the fundamentals: about 1,400 annual rows against roughly 500 companies raises an unresolved question - whether every ticker filed in all five fiscal years or whether coverage thins in earlier years - which we verify against your requested universe before delivery rather than promising blind.

None of these are hidden defects; all three are stated on the card itself. They are also precisely why the corpus works as a teaching and prototyping standard: the traps are documented, and your sample comes cut to the tickers and years you actually need.

Four complements cover what a frozen snapshot cannot:

• U.S. Treasury Interest Rate Data (Daily Yield Curve) - the discount-rate leg for any DCF you build on these fundamentals, with yield history reaching well beyond the 2016 freeze.
• Federal Reserve Bank of New York Markets Datasets - money-market and repo context for the same calendar years and beyond.
• WhaleWisdom 13F Aggregator - who held each of these tickers quarter by quarter; prices say where they went, 13F says who was positioned.
• Jay Ritter IPO Data - explains the short price histories: recently listed names like WLTW simply had fewer sessions in the window.

Browse the full set on the best investment banking & brokerage datasets ranking or the investment banking & brokerage data hub. For terminology, start with OHLCV, daily OHLCV history and ticker-to-CIK mapping.