Predicting Stock Volatility using Python

Note: Page still under construction.

In this project, I created a program that pulls historical stock prices data from AlphaVantageAPI, store it in a sqlite database, train an ARCH model to predict volatility and deploy the model via FastAPI. This is inspired by my project in WorldQuant Applied Data Science program.

API guide

POST/fit

Description: This trains an ARCH model and save it in the /models folder.

Parameters:

• ticker: a stock ticker
• n_observations: the number of past observations to train the model
• p,q: parameters of the ARCH model

Returns: JSON message with name of the trained model.

POST/predict

Gets predictions from the model.

Parameters:

• ticker: a stock ticker
• n_days: number of days to predict
• use_model: which model to use ("latest" for the latest model in the /models folder)

Returns: Predicted volatility values

Project Demo