WATER FOR ALL IN TANZANIA

Business Overview

Tanzania faces a water crisis among its population Approximately 70% of the population in Tanzania depend on water sources such as water wells . However, many of these wells are in dire need of repair, and some have even failed entirely, thus the water crisis.

Problem statement

Water shortage is a serious issue in Tanzania. It is therefore crucial to identify the category in which Water wells in the country are for repair either by the Ministry of Water in Tanzania, NGOs or Private Companies. Using Data from Ministry of Water in Tanzania, we will help categorize the Condition of the Water wells in the country to help them achieve their Goal

Objective

To build a classifier model for water wells in Tanzania

Metric of sucess

The model to have an accuracy of atleast 70%

Business Questions

1 - Who are the top 10 financiers of Water Well repairs in Tanzania

2 - Who are the top 5 Installers of water wells in Tanzania

3 - What is the quality of the Water in Tanzania

4- What is the common water source for Residents

5- What is the water basin from which water wells get water

6- What is the water wells number per region

Tools Used

python

pandas library

matplot lib

seaborn

skitlearn

K-nearest model

Modelling

The KNN model was the best model with an accuracy of 72%

Recommendation and Conclusion

Funders to consult

*The Government of Tanzania, Danida, Hesawa and world bank should be approached for financing of repairs of water wells

Installers to consult

DWE have installed the majority of water wells are best placed for reparing the water wells

Water Quality

A large portion of the water in Tanzania is Soft

Water Source

Spring water is the most common water source followed by Shallow well, machine dbh , river water and rain water harvesting

Water Basin

Pangani is the source water basin for most wells followed by Lake Victoria and Rufiji

Wells per region

Iringa region has the most water wells with over 5000 wells followed by shinyanga region with slightly 4000 water wells , kilmanajaro with slightly 4000 water wells and Morogoro region with slighly below 4000 water wells

Payment per status group

Waterpoint per status group

*The KNN model can be used as it had an accuracy of 72%

Evaluation

### Limitation

*More paramater tuning can be done on the knn model as adding more parameters affected the run time