joshuarwanda/tanzania-wells-classification

Classification of water wells in Tanzania. # Tanzania Water Wells Classification ## Problem Overview The Tanzania Ministry of Water along with Taarifa, a crowd-source platform, have commisioned the development of a predictive model that is supposed to be able to predict with **water wells** are likely to fail. While much of Tanzanias population has access to basic water services, a large 39% of households still lack this basic need. An estimated 10% of preventable deaths in the country can be attributed to inadequate *wash services*. A predictive model can enable quick **predictive maintenance** on water wells and help ensure water security in many of the rural communities that are disporportionately affected by this problem. ## Project Objectives The main objective of this project undertaking is to build a Classification model that can be able to classify water wells in Tanzania as `functional` or those that need repairs `need_repair` > **Specific Objectives** 1. To conduct exploratory analysis and determine which features to include in our model 2. Determine the cleaning steps to be included in building the model pipeline 3. To build a classifiction model that can predict the status of wells with acceptable accuracy. > **Success Metrics** * `Accuracy`: 75% * `Recall`: 80% ## The Data The data is provided by an organization known as Taarifa in co-operation with the Tanzanian government. \ A detailed description can be found here > **Column Summary** * amount_tsh - Total static head (amount water available to waterpoint) * date_recorded - The date the row was entered * funder - Who funded the well * gps_height - Altitude of the well * installer - Organization that installed the well * longitude - GPS coordinate * latitude - GPS coordinate * wpt_name - Name of the waterpoint if there is one * num_private - * basin - Geographic water basin * subvillage - Geographic location * region - Geographic location * region_code - Geographic location (coded) * district_code - Geographic location (coded) * lg …