chibui191/tanzania_water_wells

Multiple models based on Random Forest, a powerful ensemble method, to perform ternary classification of the functionality of water wells in Tanzania # Tanzania Waterpoint Functionality Classification using Random Forest Author: **Chi Bui** ## This Repository ### Repository Directory ``` ├── README.md <-- Main README file explaining the project's business case, │ methodology, and findings │ ├── data <-- Data in CSV format │ ├── notebook <-- Jupyter Notebooks for exploration and presentation │ ├── exploratory <-- Unpolished exploratory data analysis (EDA) and modeling notebooks │ ├── tanzania_cities <-- External data processing notebook │ └── report <-- Polished final notebook │ ├── predictions <-- Generated predictions on the testing dataset (for competition submission) │ ├── reports <-- Generated analysis (including presentation.pdf) │ └── images <-- Generated graphics and figures to be used in reporting ``` ### Quick Links 1. Final Analysis Notebook 2. Presentation Slides ## Overview Tanzania is in the midst of a water crisis: 4 million people in the country do not have access to a source of safe water, and 30 million people lack access to improved sanitation. People living in such circumstances, particulary women and children, usually bear the responsibility of collecting water in their communities, spending a significant amount of time traveling long distances to collect water several trips each day. The objective of this project is to use **Random Forest**, a powerful ensemble method to perform a **ternary classification** of the functionality of the water wells in Tanzania. ## Business Problems Almost half the population of Tanzania is without basic access to safe water. Although there are many waterpoints already established in the country, a lot of them are in need of repair while others have failed altogether. The model built in this project could be utilized as one of the first steps in the waterpoint functionality diagnostic process. It c …