FactorFactory R Library

Please submit your proposal to GSoC 2024 here: (FactorFactory: GSoC 2024) to improve the library. If your proposal is selected, students can get a stipend over the summer.

Introduction

FactorFactory is an R library designed to replicate financial factors documented by (Bryan Kelly) and (Chen and Zimmermann) library.

Related Work

While many R packages cater to a broad spectrum of financial analysis and factor modeling, FactorFactory focuses on providing a specialized set of functionalities tailored for the precise process of financial factor replication.

Core Functions

• Factor Construction: Develop functions to create well-known financial factors (e.g., Fama-French three-factor model) from the ground up.
• Data Management: Create utilities for efficient data preprocessing, normalization, and alignment, specifically designed for financial time series data.
• Statistical Analysis: Implement robust statistical frameworks for empirically testing factor models. This includes backtesting methodologies (e.g., using the R packages FactorAnalytics and ExpectedReturns), significance evaluations, and comparative performance analysis.

GSoC 2024 Timeline

• June 1: Introduction to factor replication
• June 7: Create R factors from Chen and Zimmermann's library
• June 14: Start with a small universe: check volume, splits and end_of_day values (e.g., dolvol and returns) and check your factors
• June 21: Larger universe robust tests
• July 01: Restructure and modularize naming conventions for different providers (e.g., eodhd, Bloomberg prioritized)
• July 07: Documentation for all data providers
• July 14: Creation of vignettes
• July 21: Integration with FactorAnalytics + Vignettes
• July 28: Integration with ExpectedReturns + Vignettes
• August 05: Creation of R CRAN package and pass all tests

Tests

Commits every 2 days to indicate project progress.

Easy

• Document features in Chen’s library not available in Bryan Kelly’s.

Medium

• Code some factors from Chen’s library not in Bryan Kelly’s documentation.

Hard

• Generate output for Chen and Zimmerman's factors.
• Establish a solid framework to ensure FactorFactory features correspond to those of Chen and Zimmerman's (unit test).

How to run the code

To run the Factor Factory script, use the following bash commands:

On Linux (please send me an email if you see other bash dependencies on Linux), first

sudo apt-get install libxml2 libxml2-dev libharfbuzz-dev libfribidi-dev libfreetype6-dev libpng-dev libtiff5-dev libjpeg-dev

Then for Windows or Linux:

git clone https://github.com/cdldl/FactorFactory.git
R -f factor_factory_run.R

Output:

[1] "94.33% of features passed the test"

Features to check:

• sale_nwc
• xido_at
• ncol_gr1a
• f_score
• cop_bev
• gp_bev
• ol_gr1a
• noa_at
• txp_gr1a
• pstk_gr1
• nri_at
• ol_gr3a
• lti_gr1a
• lti_gr3a
• div_ni
• rd5_at
• ni_ivol