GitHub Finance Data Repositories Worth Knowing For analysts, quants, and builders, the hard part is often not writing the model.

GitHub Finance Data Repositories Worth Knowing

For analysts, quants, and builders, the hard part is often not writing the model. It is finding usable data, understanding its format, and keeping the pipeline reproducible. The following GitHub repositories and datasets are useful starting points for market data, macro data, financial statements, news datasets, and asset-price pipelines.

1. Stock-market data

investment_data — A-share data, Qlib compatible

Repository: https://github.com/chenditc/investment_data

Data formats: qlib_bin.tar.gz for Qlib, plus CSV.

Data coverage: China A-share daily and minute-level bars, financial indicators, industry classifications, and suspension data.

Update frequency: Designed for daily updates, with historical data going back to around 2000.

Use case: Quantitative research and backtesting with Microsoft Qlib. The dataset is already cleaned and organized for Qlib workflows, which can save time for researchers who do not want to build the entire data-ingestion process from scratch.

How to use it: Download pre-packaged datasets from the release page or use the repository scripts to fetch and update data.

TrendRadar — Multi-platform trends plus market information

Repository: https://github.com/sansan0/TrendRadar

Data formats: JSON, Markdown, and optional Parquet.

Data coverage: Trending topics from platforms such as Weibo and Zhihu, stock quotes, financial news, and other cross-platform information streams.

Update frequency: Customizable, with schedules as frequent as hourly depending on deployment.

Use case: Monitoring public attention, market narratives, and cross-platform momentum. It can also be connected to automation workflows, notifications, or AI analysis pipelines.

How to use it: Deploy the project to generate local data files or access information through its API-style output.

FinanceDatabase — Global financial-instrument directory

Repository: https://github.com/JerBouma/FinanceDatabase

Data formats: CSV and JSON.

Data coverage: More than 300,000 financial instruments, including stocks, ETFs, funds, indices, cryptocurrencies, and other market symbols.

Update frequency: Updated periodically.

Use case: Symbol mapping, instrument classification, universe construction, metadata enrichment, and connecting tickers to other market-data sources.

Important note: It is not a real-time quote source. Its value is in instrument coverage and metadata.

How to use it: Clone the repository and use the data files directly.

2. Macro-economic data

Global-Macro-Database

Repository: https://github.com/KMueller-Lab/Global-Macro-Database-R

Data formats: CSV and RData.

Data coverage: Hundreds of countries and dozens of macro variables, including GDP, inflation, interest rates, trade, and other long-term indicators. The dataset combines modern and historical sources.

Update frequency: Annual releases.

Use case: Long-horizon macro research, cross-country comparisons, economic-history analysis, factor research, and regime studies.

How to use it: Download the full dataset from the release page or access it through the R package workflow.

FirstData — Integrated international-organization datasets

Repository: https://github.com/MLT-OSS/FirstData

Data formats: CSV, JSON, Excel, and SDMX.

Data coverage: World Bank WDI, IMF data, OECD statistics, United Nations data, and other international-organization datasets.

Update frequency: Designed to synchronize with releases from international organizations.

Use case: Building standardized macro data pipelines, retrieving country-level indicators, and creating reproducible economic dashboards.

How to use it: Clone the repository and run scripts to fetch the selected datasets.

3. Financial-statement and company-news data

sec-financial-statement-data-set

Repository: https://github.com/HansjoergW/sec-financial-statement-data-set

Data formats: Parquet and CSV.

Data coverage: U.S. SEC company filings, including 10-K and 10-Q financial statements such as balance sheets, income statements, and cash-flow statements.

Update frequency: Quarterly updates aligned with SEC EDGAR data releases.

Use case: Fundamental research, factor construction, financial-statement screening, accounting-based models, and cross-company analysis.

How to use it: Use the scripts to download SEC data and convert it into Parquet for efficient querying.

FNSPID Financial News Dataset

Repository: https://github.com/Zdong104/FNSPID_Financial_News_Dataset

Data formats: CSV and Parquet, with storage linked through Hugging Face.

Data coverage: A large-scale combination of stock prices and financial news, covering millions of news records and price observations across thousands of U.S. companies.

Use case: Natural-language-processing research, sentiment analysis, news-driven market studies, event studies, and AI model training for finance.

How to use it: Access the dataset through Hugging Face Datasets or links provided in the repository.

4. Crypto, derivatives, and asset-price pipelines

Financial-Data-Pipeline

Repository: https://github.com/AhmedAli58/Financial-Data-Pipeline

Data formats: Parquet, CSV, and JSON.

Data coverage: Cryptocurrency, stock, and derivative-market data, with historical and real-time components depending on configuration.

Update frequency: Real-time or scheduled updates, with support for incremental synchronization.

Use case: Modular ETL, data-quality monitoring, lineage tracking, and multi-asset research pipelines.

How to use it: Deploy the pipeline and generate local files or connect through the project’s API-style workflow.

asset-prices-parquet-saver

Repository: https://github.com/JeremyWhittaker/asset-prices-parquet-saver

Data format: Parquet.

Data coverage: Historical prices for stocks, ETFs, and cryptocurrencies from sources such as Alpaca and Yahoo Finance.

Update frequency: Customizable, with incremental update support.

Use case: Backtesting, research notebooks, analytics dashboards, and local data lakes where Parquet is preferred for speed and compression.

How to use it: Run the scripts to download and store asset prices as Parquet files.

The main lesson: choose the dataset based on the research question. For market backtesting, use clean price and corporate-action-aware data. For macro research, prioritize source consistency. For financial statements, prefer structured SEC-derived datasets. For AI and NLP, look for large news datasets with timestamps and labels.

For production workflows, prefer formats such as Parquet and pipelines with incremental updates.

Educational content only. Not investment advice.