amaliza-shal/PCA-African-Financial-Inclusion--Formative2
Domain:
socioeconomicRecord type:
software
Creator:
ama
Host:
PCA: Financial Inclusion in East Africa.
What is this?
This project is a deep dive into who has access to banking in Kenya, Rwanda, Tanzania, and Uganda. I used Principal Component Analysis (PCA) to take a complex list of 33 demographic features and boil them down to the ones that actually matter.
Key Features
From-Scratch PCA: Built the math (Covariance, Eigendecomposition) without using high-level libraries like scikit-learn.
Smart Selection: The code automatically picks enough components to keep 95% of the original data's information.
Performance: Optimized to process over 23,000 rows of African survey data in milliseconds.
Setup & Usage
Install: pip install -r requirements.txt
The Library: I’ve included pca_lib.py, which is my own custom tool for this analysis.
Run it:
Python
from pca_lib import PCA_Analysis_Tool
pca = PCA_Analysis_Tool(n_components=2)
pca.fit(X_standardized)
reduced_data = pca.transform(X_standardized)
How to finalize your submission:
Download from Colab:
The Notebook: File -> Download -> .ipynb.
The Library: Click the folder icon on the left of Colab, right-click pca_lib.py, and hit Download.
GitHub Upload:
Go to your GitHub repo.
Click Add file -> Upload files.
Drag in your Notebook, pca_lib.py, requirements.txt, and Train.csv.
The README:
Open the README.md on GitHub, click the pencil icon, and paste the "Humanized" text above.
Save/Commit.