Revised 23 October 2026 · v1.0.0 · what changed

Code: github.com/herrrickshaw/colab-market-fundamentals

Watch the walkthrough

Watch on YouTube: youtu.be/9NjzvuAPlfE

Why it is useful

• Who benefits: students and researchers who want a table of annual revenue, profit, assets, debt and cash flow for many companies without paying for a data feed.
• What it saves: building a rate-limited, resumable collector yourself. The notebook already handles Yahoo blocking, retries and checkpoints.
• Evidence from the run: a probe of three large companies passed, and a live test of the collector on three German companies (SAP, Siemens, Allianz) returned 12 rows with 13 columns in a few seconds.
• It stores one row per company per fiscal year with the fields needed for a Piotroski-style score, so the output slots straight into a screening notebook.

What it is for

Open collect.ipynb in Colab, press Run all, and download fundamentals.zip. The notebook reads tickers.csv, skips the US by default and the India set is stored with bare symbols, fetches income statement, balance sheet and cash flow for each ticker, and writes one parquet file per market.

The data come from Yahoo Finance through yfinance, so they are third-party and unaudited. Use them for research and teaching. Not investment advice: this is research and education material built from public-source or derived data; it is not a recommendation to buy or sell anything.

Who it is for

• Students and analysts doing multi-market fundamentals research
• People who want a reproducible, resumable collection job that runs in the free Colab tier
• Non-programmers who want an AI assistant to walk through each cell

What you need

Get the code and an AI assistant

You do not have to be a programmer. Get the repository, give it to an AI assistant, and ask in plain language. Any of these routes works:

git clone https://github.com/herrrickshaw/colab-market-fundamentals.git
cd colab-market-fundamentals
claude      # or: gemini

Install

1. Open in Colab

   https://colab.research.google.com/github/herrrickshaw/colab-market-fundamentals/blob/main/collect.ipynb

2. Or run locally in a virtualenv

   python3 -m venv .venv && source .venv/bin/activate
   pip install yfinance curl_cffi pandas pyarrow
   git clone https://github.com/herrrickshaw/colab-market-fundamentals.git

Quick start: commands and what you should see

Cell 1: install and load the ticker universe

U = pd.read_csv('tickers.csv')
print(U.groupby('market').size().to_dict())

The universe size (32,590 tickers, 20 markets) and a count per market.

Cell 4: probe, then collect (or test one market first)

collect_market('DE', ['SAP.DE','SIE.DE','ALV.DE'], workers=3)

'DE: 3 names with data'. A full Run all prints progress per market and writes /content/fundamentals.zip every three minutes.

Sample outputs

Real output from a run on the date shown in the sources line; trimmed to fit.

Universe by market (cell 1, run locally from the committed tickers.csv)

universe: 32590 tickers, 20 markets
{'AU': 1509, 'BR': 243, 'CA': 2091, 'CH': 192, 'CN': 2895, 'DE': 449, 'DK': 104, 'EU': 1067, 'FI': 171, 'HK': 2793, 'IN': 2368, 'JP': 3709, 'KR': 2768, 'SA': 374, 'SE': 564, 'SG': 614, 'TW': 2204, 'UK': 854, 'US': 7441, 'ZA': 180}

The US set is skipped by default in the collection loop.

IP probe and a three-company test (collector code from the notebook, run locally)

probe 3
  DE: 3 names, 0 done, 3 to fetch
  DE: 3 names with data
(12, 13)

Three large companies, four fiscal years each, 13 columns.

Latest two fiscal years, selected columns (EUR)

ticker  fy_end      net_income    revenue       cfo           filed
SAP.DE  2024-12-31  3.124e9       3.4176e10     5.207e9       2025-03-31
SAP.DE  2025-12-31  7.161e9       3.680e10      9.156e9       2026-03-31
SIE.DE  2024-09-30  8.301e9       7.5930e10     1.1665e10     2024-12-29
SIE.DE  2025-09-30  9.620e9       7.8914e10     1.3257e10     2025-12-29

Yahoo-sourced figures, shown only to illustrate the table shape; the 'filed' column is fiscal year end plus 90 days.

Columns written per company-year

ticker, fy_end, net_income, revenue, gross_profit, total_assets, current_assets,
current_liabilities, long_term_debt, shares, equity, cfo, filed

Some fields are missing for some companies, for example banks and insurers have no gross profit or current assets.

Ways to use it

Sample prompts for your AI assistant

Copy, paste and edit the parts in [brackets].

• Claude: Read collect.ipynb and explain in plain English how the checkpointing and the IP probe work.
• Gemini: Modify cell 4 so only the DE and SE markets run, and print how many rows each parquet has at the end.
• ChatGPT: I uploaded a market parquet from this notebook. Compute return on equity per company-year and flag missing values.
• GitHub (Copilot Chat): Add a field for 'Total Debt' to the BAL dictionary in the collector and update the output columns.
• Claude: Explain why the 'filed' date here cannot be used for a point-in-time backtest and suggest a safer approach.

Limits and honest caveats

Sources: Ran the notebook's cell 1 logic and collector cell 3 (with probe) locally in a throwaway virtualenv on 19 October 2026 from a clone of github.com/herrrickshaw/colab-market-fundamentals, using yfinance with curl_cffi for three large German companies. Colab-specific cells (upload, download) not run.

About this article: Researched, written and edited by Umashankar Triplicane Dwarakanathan, with AI research assistance; every figure is meant to trace to the primary source cited. See the Editorial Policy for how sourcing, AI use and corrections work.