jsiddiqui85/Tanzania_Water_Wells_Predictive_Analysis

Using classification models to analyze and predict on the Tanzania Water Wells. These models will assist the government of Tanzania to predict which water wells need to be repaired before they become unusable. # Tanzania Water Wells (Classification Models) **Author:** Jawwad A. Siddiqui Data Scientist github.com linkedin.com # Overview This project uses classification to model and predict the state of water wells in the African nation of Tanzania. The purpose for the predictive model is to assist the government officials at `The Tanzania Commission for Water` to find water wells that are `not functioning at all` or that are `functioning but require minor/major repairs` so the agency can subcontract the work out to bring these water wells up to functioning level. I have used a curated data set found on `DrivenData.com` to run my predictive models against. Prior to cleaning the data, this dataset contained over `60k rows` along with `30+ features`. # Business Understanding `The Tanzania Commission for Water` is a federal agency in the African State of Tanzania primarily responsible for the cleanliness and safety of the water that is supplied to the general public and residents of Tanzania. Currently, Tanzania lacks the ability to provide a safe source of drinking water to over `50% of their population`. Although there have been many attempts to correct this issue over the years, so far all of those attempts have fallen short of their objectives - leaving `45% of their water wells in need of major repairs`. My goal with this project is to help `The Tanzania Commission for Water` and the director of this agency with identifying which wells need to be repaired currently, while also predicting on which of these wells will need to be repaired in the future in order to maintain the cleanliness and safety of the water that is supplied to the Tanzania population. # Data My model and recommendations come from work done on the `Tanzanian Water Wells` data set found on `Driven Data`. From this dataset, I looked at various features that would best predict the water wells that need to be repaired. My best model is able to …