The purpose of this project is to understand if we can use data collected by Tennessee fuel inspectors to better predict gas station fuel test failures?

• Inferential Statistics
• Machine Learning
• Data Visualization
• Predictive Modeling
• Time Series Analysis

• Python
• Pandas
• jupyter

In this project, we are interested in understanding if we can predict when a fuel compliance test will fail.

The state of Tennessee's Department of Agriculture (TDA) maintaines a fuel quality inspection program. Each year, the state inspects all places where fuel is sold/ distributed including gas stations, terminals and airports. The results of these routine tests as well as follow up tests for complaints are maintained by the Tennessee Department of Agriculture.

Although around 97% of the tests pass these compliance tests, each year around 3% of the tests fail. A failing test could mean that consumers are exposed to fuel that could do harm to their property or themselves.

1. Clone repo https://github.com/99KHayes/tenessee-fuel-quality-analysis.git
2. Raw Data is being kept in the data/01_raw folder within this repo.
3. The final results technical notebook can be found here. The notebook is located in the results folder.
4. If you would like to run the code please run the jupyter notebook found in the results folder.
5. Data processing/transformation scripts are being kept in src NOTE: We have included only a sample of our dataset (results data from 2016) and the ASTM set.

• Executive Summary Slidedeck
• Technical Notebook

This template is based on the DSSG machine learning pipeline.