terence-lim / financial-data-science

Financial and Investment Data Science: FinDS Python library and examples for applying quantitative and machine learning methods on structured and unstructured financial data sets

FINANCIAL DATA SCIENCE

FinDS Python package for financial data science projects

• use database engines SQL, MongoDB, Redis
• interfaces for
  • structured data from CRSP, Compustat, IBES, TAQ
  • APIs from ALFRED, BEA
  • unstructured data from SEC Edgar, Federal Reserve websites
  • academic websites by Ken French, Loughran and MacDonald, Hoberg and Phillips
• recipes for econometrics, finance, graphs, event studies, backtesting
• applications of statistics, machine learning, neural networks and large language models

Resources

1. Online Jupyter-book, or download pdf

2. FinDS API reference

3. FinDS repo

4. Jupyter notebooks repo

Examples

notebook	Financial	Data	Science
stock_prices	Stock distributions, delistings	CRSP stocks	Statistical moments
jegadeesh_titman	Overlapping portfolios; Momentum effect	CRSP stocks	Hypothesis testing; Newey-West estimator
fama_french	Portfolio sorts; Value effect	CRSP stocks; Compustat	Linear regression;
fama_macbeth	Cross-sectional Regressions; CAPM	Ken French research library	Non-linear regression; Quadratic optimization
weekly_reversals	Mean reversion; Implementation shortfall	CRSP stocks	Structural breaks; Performance evaluation
quant_factors	Factor investing; Backtests	CRSP stocks; Compustat; IBES	Cluster analysis
event_study	Event studies	S&P key developments	Multiple testing; FFT
economic_releases	Economic data revisions; Employment payrolls	ALFRED	Outliers
regression_diagnostics	Consumer and producer prices	FRED	Linear regression diagnostics; Residual analysis
econometric_forecast	Production and Inflation	FRED	Time series analysis
approximate_factors	Approximate factor models	FRED-MD	Unit root test
economic_states	State space models	FRED-MD	Gaussian Mixture; HMM
term_structure	Interest rates	FRED yield curve	SVD
bond_returns	Bond risk factors	FRED bond returns	PCA
option_pricing	Binomial tree; Black-Scholes-Merton and the Greeks	Simulations	Monte Carlo simulation
conditional_volatility	Value at risk	FRED crypto-currencies	EWMA; GARCH
covariance_matrix	Portfolio risk	Fama-French industries	Covariance matrix estimation
market_microstructure	Market impact; Liquidity risk	TAQ tick data	High frequency volatility
event_risk	Earnings misses	IBES	Poisson regression; GLM
customer_ego	Supply chain	Compustat principal customers	Graph networks
industry_community	Industry sectors	Hoberg and Phillips research library	Community detection
bea_centrality	Input-output tables	Bureau of Economic Analysis	Graph centrality
link_prediction	Product markets	Hoberg and Phillips	Link prediction
spatial_regression	Earnings surprises	IBES Hoberg and Phillips	Spatial regression
fomc_topics	FOMC meetings	Federal Reserve website	Topic models
mda_sentiment	10-K filings	SEC Edgar; Loughran and Macdonald research library	Sentiment analysis
business_description	10-K filings	SEC Edgar	POS tagging; Density-based clustering
classification_models	Industry classification	SEC Edgar	Classification
regression_models	Macroeconomic forecasts	FRED-MD	Regression
deep_classifier	Industry classification	SEC Edgar	Neural networks; Word embeddings
recurrent_net	Macroeconomic models	FRED-MD	RNN; Dynamic factor models
convolutional_net	Macroeconomic forecasts	FRED-MD	CNN; Vector autoregression
reinforcement_learning	Spending policy	SBBI	Reinforcement learning
fomc_language	Fedspeak	FOMC meetings minutes	Language modelling; Transformers
sentiment_llm	Financial news sentiment	Kaggle	LLM prompt engineering
summarization_llm	10-K filings	SEC Edgar	LLM text summarization
finetune_llm	Industry classification	SEC Edgar	LLM fine-tuning
rag_agent	Corporate philanthropy	textbooks	LLM RAG, chatbots, agents

Contact

Github: https://terence-lim.github.io