mduncan23/Tanzanian-Water-Wells-Analysis

Using supervised machine learning to predict the repair status of wells located across the country of Tanzania. # Tanzanian Water Wells **Author**: Matthew Duncan ## Overview For this analysis, I have analyzed data on almost 60,000 wells across the country of Tanzania. I've utilized descriptive analysis, visualizations, and machine learning methods to provide recommendations on focus points for well repair efforts. ## Business Problem Tanzania struggles with providing clean water to the people of its country. The goal of this analysis is to use data from existing wells throughout the country to identify the wells in need of repair. This analysis is focused on assisting the Tanzanian Government reduce resources spent to identify water wells that are in need of repair to ensure that clean water is available to its citizens. ## Data The data was provided by DrivenData.org in partnership with Taarifa who aggregated the data from the Tanzanian Ministry of Water. ## Methods This analysis uses descriptive analysis, visualizations, and machine learning methods to describe trends for wells in need of repair across the country. Additionally I have provided recommendations on focus points for well repair efforts To better understand the data, I have reviewed the features and separated them into numerical and categorical feature types. To better define the status of each well for this analysis, I have also set up a binary system for the wells: - 0 = Functional - 1 = Needs Repair About 54% of all of the wells are completely functional and 46% of the wells need repairing. To ensure that I'm reducing the noise in the dataset while simultaneously being cognizant of computing time, I've removed a few columns that are not relevant to this analysis or are duplicated in other features: - `scheme_name` - Who operates the waterpoint. - `id` - A unique identifier number for each well. - `date_recorded` - The date the row was entered. - `funder` - Who funded the well. - `recorded_by` - Group entering this row of data. - `wpt_name` - Name of the waterpoint if there is one. - `regio …