This is an example project of retrieving data from twelvedata and using Monument to predict stock returns.

Overview

In this project, we predict the weekly return for the Amazon stock AMZN one week ahead, by using the Gaussian Process-integrated LightGBM algorithm (G-Boost). We use historical prices and a few technical indicators as independent variables. To evaluate model performance, we use directional accuracy as the metric, where we achieved 56% in validation and 57% in serving.

For more details, please read mai_twelvedata_project.ipynb.

Developing Environment

• Python 3.9.2

  • pandas 1.2.3

  • requests 2.25.1

• Monument.app 1.4.13

  • monument (python package) 0.0.1

Further Reading

• Monument quickstart: https://monument.ai/m/quick-start

• Monument Python package: https://pypi.org/project/monument

• twelvedata documentation: https://twelvedata.com/docs#getting-started

• getting twelvedata API key: https://twelvedata.com/register