Final project from Wesleyan University's "Data Analysis and Interpretation" specialization on Coursera / DrivenData.org project. Repository includes predcitive model and written report of findings.
-
Classification Model: https://github.com/mvenouziou/Project-Exploratory-Data-Analysis-and-Classification/blob/main/water_table_Driven_Data.ipynb
Clean, potable water is a basic human need. Beyond mere access to water, proximity to a reliable water source is a major factor in determining the educational and economic opportunities available to adjacent communities. Because of this, the development and maintenance of water sources in remote and/or impoverished communities is a major area of focus for many government and nonprofit humanitarian organizations.
This paper develops a predictive model for determining the maintenance needs of water pumps in Tanzania. Specifically, this is a supervised learning problem to classify pumps as functional, in need of repair or completely inoperable. The model achieves 79.63% predictive accuracy on the DrivenData test set using decision tree models.