NathanielMartin42/Tanzania-Well-Project

A project to classify well condition in Tanzania # Tanzania Well Project Tanzania is located in Eastern Africa with a population of about 59,678,000 people. Access to water remains a major challenge in Tanzania especially in its rural villages. The hope is that by predicting the functional status of wells, Tanzania can maintain the existing systems and identify the neglected wells as well as the development of new water delivery mechanisms. ## Business Problem The purpose of this project is to use Machine Learning Classification models to predict the functional status of water wells in Tanzania. These models include Decision Trees, Random Forest, and Gradient Boost using RandomizedSearchCV to identify the best parameters for each. The different status groups for classification are functional, non functional, and functional but needs repair. The hope is that by gaining a better understanding of the factors that impact water wells, we are able to improve maintenance and ensure all wells stay functional. The importance of having accessibility of water for these Tanzanian communities is that it decreases diseases, increases education, and increases the economy overall. ## Data The data that was used to train our models was provided by Taarifa and the Tanzanian Ministry of Water. The data sets are as follow: - Training-set-values: 59,400 observations, 40 variables - Training-set-labels: 59,400 observations; contains status group labels We perform a train test split on this set in order to properly test our models on previously unseen data. - Test-set: 14,850 observations, 40 variables This is provided by the competition website to generate predictions on, and rank the outputs. You can download the datasets here. Additionaly, we used SQL to join the Training-set-values and Training-set-labels to make the Joined_values_labels csv to make plotting easier on Tableau. ## Pipeline We created a Class to organize our modeling process and allow quick use of our chosen pipeline objects. This class performs our train …