Recommended libraries

Grouped after the Awesome AI4Finance taxonomy, expanded with our Tech Stack and blog. Scores mirror that list’s “Recommendation” column (community curation, not performance advice).

Awesome AI4Finance — Structure, categories, and 1–5 scores follow the community-maintained list below; we added an extra section for infrastructure and research tools we reference on MarketMaker.cc.

Score 1–5 = strength of signal in the Awesome list (weekly review process described there). GitHub star counts are indicative—check each repo for live numbers.

Financial Big Data

Data sources

Project Stars Score Description
FinRL-Meta 1.8k ★★★★★ Metaverse of market simulators for DRL—stocks, crypto, FX, paper and live connectors.
CCXT 41.8k ★★★★★ Unified REST/WebSocket exchange APIs in JS/Python/PHP—baseline for many crypto connectivity prototypes.
StockSharp 9.7k ★★★★ C#-centric stack for stocks, FX, crypto, and options with strategy tooling.
TuShare 14.7k ★★★ China A-share and macro data access—useful when your universe includes mainland listings.
yfinance 22.9k ★★★ Quick Yahoo Finance historical series—great for sketches, not a vendor SLA.
Binance public APIs 3.0k ★★★ Reference implementations and API docs for one of the deepest crypto liquidity pools.
Alpaca (Marketstore) 1.7k ★★★ Columnar market datastore plus brokerage APIs for equities/crypto paper and live workflows.
WRDS 155 ★★ Academic data access patterns for Wharton Research Data Services subscribers.

Features & technical indicators

Project Stars Score Description
TA-Lib 11.9k ★★★★★ Industry-standard technical indicator primitives for feature pipelines.
Clairvoyant 2.4k ★★★ Social/historical cue monitoring for short-horizon equity experiments.
FinanceDatabase 7.4k ★★★ Symbol metadata across equities, ETFs, funds, FX, crypto, and money markets.

Artificial Intelligence

Machine learning

Project Stars Score Description
Machine Learning for Trading (book repo) 17k ★★★★★ Companion code to Stefan Jansen’s book—end-to-end ML workflows for trading research.
Qlib 40.6k ★★★★ Microsoft’s AI-first quant stack: data, models, and experimentation in one toolkit.
Stock-Prediction-Models 9.3k ★★★★ Large zoo of classical and deep forecasting baselines for benchmarking.
TF Quant Finance 5.3k ★★★ Google’s TensorFlow primitives for derivatives pricing and simulations.
Adv_Fin_ML_Exercises 1.9k ★★★ Worked solutions tied to Marcos López de Prado’s Advances in Financial Machine Learning.
AlphaPy 1.7k ★★★ Feature engineering + modeling utilities aimed at discretionary and quant researchers.
fin-ml 1.2k ★★★ Case-study code for “Machine Learning and Data Science Blueprints for Finance.”

Reinforcement learning

Project Stars Score Description
FinRL 14.7k ★★★★★ Full DRL pipeline for finance—data, environments, training, and deployment stories.
ElegantRL 4.3k ★★★★★ Scalable PyTorch RL algorithms with cloud-friendly elasticity patterns.
TensorTrade 6.1k ★★★★ Modular RL framework for training and evaluating trading agents.
FinRL-Trading (ICAIF ensemble) 1.3k ★★★★ Ensemble DRL strategies with a published live-trading experiment path.
gym-anytrading 2.4k ★★★ Minimal Gym environments for price-series RL tutorials.

Finance

Stock recommendation

Project Stars Score Description
Machine Learning for Stock Recommendation (IEEE) 34 ★★ Reference implementation for a classical supervised recommendation baseline.

Trading

Project Stars Score Description
HFT-LOB-Trading-ML 2.3k ★★★ Full order-book tick ML baselines—useful when studying microstructure-heavy datasets.

Portfolio management

Project Stars Score Description
PyPortfolioOpt 3.2k ★★★★ Mean-variance, Black–Litterman, HRP, and related optimizers in Python.
OLPS 358 ★★ Online portfolio selection algorithms for sequential decision benchmarks.

High performance computing

Project Stars Score Description
NumPy 31.8k ★★★★★ Foundational ndarray stack for virtually every Python quant library.

Trading platforms

Project Stars Score Description
QuantConnect Lean 18.4k ★★★★ Open-source algorithmic engine with research, backtest, and live brokerage adapters.

Rendering & visualization

Project Stars Score Description
TradingGym 1.9k ★★★ RL training/backtesting with richer rendering hooks for agent diagnostics.
mplfinance 4.3k ★★★ Candlestick and market charts on top of Matplotlib.

Databases, APIs, messaging, and research libraries we use in articles and on the landing—editorial picks with informal scores for orientation only.

Extended stack (MarketMaker.cc & blog)

Time-series databases & OLAP

Project Stars Score Description
QuestDB / ClickHouse / DuckDB ★★★★ Columnar engines for ticks, bars, and analytics—called out in our Tech Stack section.

APIs, RPC & streaming

Project Stars Score Description
PostgreSQL + Hasura (GraphQL) ★★★★ Typed data APIs over Postgres—matches our GraphQL/Hasura pairing on the landing.
gRPC ★★★★ Binary RPC between microservices; discussed alongside WebSocket/OpenAPI in our architecture posts.
Redis 73.8k ★★★★ Caches, pub/sub, and hot state in execution pipelines (data-communication article).
Kafka 32.4k ★★★★ Durable event streaming for market-data fan-out and async services.

Low-latency messaging

Project Stars Score Description
Aeron ★★★★★ Microsecond-class messaging—deep dive on our blog; pairs with SBE/FIX ecosystems.

Backtesting, features & research

Project Stars Score Description
vectorbt / backtesting.py 7.2k ★★★★★ Vectorized research loops; VectorBT covered in a dedicated article.
Numba 11k ★★★★ JIT acceleration for NumPy-style loops—core to fast vectorized backtests.
hmmlearn 3.4k ★★★★ HMM baselines for regime detection—used in our adaptive-trading article.
scikit-learn 65.8k ★★★★ Classical ML baselines, CV, and pipelines for tabular alpha research.
Stable-Baselines3 13.1k ★★★★ Reference RL algorithms (PPO, SAC, …) for research prototypes.
FinGPT / FinNLP (ecosystem) 19.1k ★★★★ Open financial LLM/NLP lines for sentiment and document prototypes.

Educational overview. Third-party names and scores do not imply endorsement; verify licenses, latency, and compliance before production use.