OnzyBoy/Financial-Inclusion-in-East-Africa
Domaine:
socioeconomicType de record:
project
Créateur:
Onz
Hôte:
# 🌍 Financial Inclusion Prediction in East Africa
## 📌 Project Overview
This project predicts the likelihood of an individual having a **bank account** in East Africa (**Kenya, Rwanda, Tanzania, and Uganda**) using demographic and socioeconomic data.
It was completed as part of the **GOMYCODE Data Science Capstone Project**.
---
## 📊 Dataset
The dataset is provided by Zindi Africa. To respect the competition rules, the data files are **not hosted in this repository**.
🔗 **Data Source:** Financial Inclusion in Afri…
---
## 🌐 Live Demo (Streamlit App)
An interactive web application was built using Streamlit to allow users to make real-time predictions.
🚀 **Try the App:**
Live Demo
💡 *Input user details and instantly see the predicted likelihood of having a bank account.*
---
## 📈 Interactive Dashboard
An interactive dashboard was created using Tableau to explore key insights visually.
🔗 **View Dashboard:**
public.tableau.com
💡 *For the best experience, open the dashboard in **fullscreen mode**.*
---
## 🔄 Methodology (CRISP-DM)
- 🧠 **Business Understanding:**
Identifying key factors that drive financial inclusion and support outreach strategies.
- 🧹 **Data Preparation:**
Data cleaning, handling missing values, One-Hot Encoding, and feature scaling.
- 🤖 **Modeling:**
Compared multiple machine learning models:
- Logistic Regression
- Decision Tree
- Random Forest
- XGBoost
- Support Vector Machine (SVM)
- Naive Bayes
- 🏆 **Champion Model:**
**Random Forest Classifier**
- Accuracy: **82%**
- Recall: **75%**
- F1-Score: **0.53**
- ROC AUC: **0.87**
---
## 🔍 Key Insights
- 🎓 **Education Level** and 📱 **Cellphone Access** were the strongest predictors of financial inclusion.
- 🎯 The model achieved a **Recall of ~74%**, meaning it successfully identified **3 out of 4 individuals** who have ban …