TDA of financial time series

This web app is a reimplementation of the paper TDA of financial time series: Landscapes of crashes using gudhi.

Interactive dashboard from notebook with Voilà

How to run it?

Voilà	JupyterLab

Binder or Voilà instance creation may take some time, but it requires no local installation.

Local installation

One can install it easily from conda:

• Install miniconda3 environment .
• Create a new conda environment and activate it:

conda env create -f environment.yml
conda activate tda_financial_time_series

Run the web app

voila tda_web_app.ipynb

jupyter lab tda_web_app.ipynb

Data generation

Data are generated using Yahoo! finance pip module to get the four major US stock market indices: S&P 500, DJIA, NASDAQ, and Russel 2000. To install Yahoo! finance pip module:

conda install pip --yes
pip install yfinance

To generate the data with the latest values (should be done by the Continuous Integration automatically):

python data_generation.py

A new version of latest.csv is generated.