# Phase_3_project Tanzania Water Wells
# Overview
The problem of providing clean water to the population of over 57,000,000 in Tanzania is a major concern, as many existing water points in the country are in need of repair or have failed altogether. To address this issue, a classifier will be built to predict the condition of a water well based on information such as the type of pump, date of installation and others. The target audience for this classifier could be an NGO focused on repairing wells or the Government of Tanzania looking to improve the construction of new wells.
# Business Understanding
The objective of building this classifier is to assist the NGO or the Government of Tanzania in their efforts to provide clean water to the population. By predicting the condition of a water well, the NGO can prioritize their resources and focus on the wells that are in need of repair, while the Government of Tanzania can use the insights to make informed decisions about the construction of new wells. The classifier will provide a reliable and efficient solution to the problem of ensuring clean water in Tanzania, which is essential for the health and well-being of its population.
# Data Understanding
The data used in this analysis was obtained from DrivenData and was sourced from Taarifa and the Tanzanian Ministry of Water . It contains information on water wells in Tanzania and is divided into three files, including training set values, training set labels, and test set values. The training data has 59,400 observations and 41 variables, providing extensive information on various aspects of the water pumps:
# Modelling
- The goal of this project was to evaluate and compare the performance of three machine learning models on a binary classification problem with imbalanced data. The models used were a baseline Logistic Regression Model, a tuned Random Forest Model, and a tuned XGBoost Model.
- The data was preprocessed to handle the imbalance issue by using …