Logo Lanfrica
  • Home
  • Atlas
  • Insights
  • Docs
  • Sign in

© 2026 Lanfrica. All rights reserved. All copyrights of the resources shown on the Lanfrica website belong to the original copyright holders, unless explicitly stated otherwise.

jonnyhof/tanzanian_waterpoint_classification

Domain:

environment and energy

Record type:

dataset
Creator:
jon
Host:
Classifying the working status of water wells in Tanzania # Tanzanian Waterpoint Functionality Classification Classifying the working status of waterpoint pumps in Tanzania. #### By Jonny Hofmeister This data comes from Taarifa, an open source platform that tracks infrastucture related issues and data, who has sourced this data from the Tanzanian Ministry of Water. It is posted on the DRIVENDATA site as a classification data science contest. ## Summary This classification and analysis in the follows the CRISP-DM process for data mining. This README provides a brief summary of the project and the results. ### *Task* In order to better understand waterpoint functionality accross the country, the Tanzanian Ministry of Water would like to build a classification model that predicts the status of waterpoint pumps with unknown functionality. With an accurate model and a more complete understanding of functionality, they can more efficiently address repairs. Further down the road, information from the model can be used help to inform them of patterns of how and why pumps fail. The image below shows the number of repairs needed for each Ward in Tanzania. Wards without data are shown in gray. ### *Data* The target in the data we would like to predict is 'status_group' and is reported in three classes - 'functional', 'non functional', and 'functional needs repair'. There is slight class imbalance in the target, functional makes up 54.4% of the rows, non functional and functional needs repair make up 38.4% and 7.2% respectively. The goal is to produce a ternary classification model that can label predict all three status groups. Around 40 features are given in the original dataset. They contain numerical information like population, gps position, and total static head, as well as categorical features like region, ward, waterpoint type, quantity, and quality. Many of these features come in pairs that are extremely similar and redundant. The features were trimmed to leave out duplicate types of information, but still encomp …

Visit

github.com

Tasks

text classification