https://github.com/fbarffmann/deep-learning-challenge
Predicted donation outcomes for a nonprofit using TensorFlow and neural networks. Includes data preprocessing, model tuning, and evaluation to optimize binary classification accuracy.
https://github.com/fbarffmann/deep-learning-challenge
Last synced: about 21 hours ago
JSON representation
Predicted donation outcomes for a nonprofit using TensorFlow and neural networks. Includes data preprocessing, model tuning, and evaluation to optimize binary classification accuracy.
- Host: GitHub
- URL: https://github.com/fbarffmann/deep-learning-challenge
- Owner: fbarffmann
- Created: 2024-09-18T21:23:00.000Z (7 months ago)
- Default Branch: main
- Last Pushed: 2025-04-01T13:13:05.000Z (23 days ago)
- Last Synced: 2025-04-01T13:22:48.384Z (23 days ago)
- Language: Jupyter Notebook
- Homepage:
- Size: 940 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Deep Learning Challenge: Charity Donation Predictor
## Overview
This project applies a supervised deep learning model to predict whether applicants will be likely donors to a charitable organization. The goal is to help optimize fundraising efforts by identifying which applicants are most worth pursuing for donations.
The dataset comes from a real-world scenario provided by AlphabetSoup, a fictitious non-profit, and includes anonymized features related to application activity and demographics.
## Objectives
- Preprocess and encode categorical variables from the charity dataset.
- Build a binary classification model using TensorFlow/Keras.
- Optimize the model to achieve a minimum accuracy of **75%**.
- Evaluate the model’s performance and analyze its limitations.
---
## Technologies Used
- **Python**
- **TensorFlow / Keras**
- **Pandas**
- **Scikit-learn**
- **Jupyter Notebook**
---
## Approach
1. **Preprocessing the Data**
- Loaded and inspected the dataset from `charity_data.csv`.
- Dropped non-informative columns (like EIN and NAME).
- Converted categorical features into numerical encodings using one-hot encoding.
- Split the data into training and testing sets.
2. **Building the Neural Network**
- Created a baseline Sequential model with:
- Input layer matching the number of features.
- Two hidden layers using ReLU activation.
- Output layer using Sigmoid activation for binary classification.
- Compiled the model using `binary_crossentropy` loss and the `adam` optimizer.
- Trained the model for 100 epochs and evaluated performance.
3. **Optimization Attempts**
- Tuned number of neurons and layers.
- Tried additional dropout layers to reduce overfitting.
- Tested with different activation functions and epoch counts.
---
## Results
- The final optimized model achieved a testing accuracy of **~75.3%**, meeting the target benchmark.
- Adding complexity (more neurons/layers) did not consistently improve performance.
- Feature engineering (e.g. excluding high-cardinality or irrelevant variables) played a bigger role in improving accuracy.
---
## Key Takeaways
- Neural networks can be effectively applied to classification problems with structured, tabular data.
- Preprocessing decisions—such as which features to encode or drop—can have as much impact as tuning hyperparameters.
- Even a relatively simple model can achieve strong performance when fed clean and relevant data.
---
## How to Use
1. Clone the repository:
```bash
git clone https://github.com/fbarffmann/deep-learning-challenge.git
```
2. Navigate to the project directory:
```bash
cd deep-learning-challenge
```
3. Install required libraries if needed:
```bash
pip install pandas scikit-learn tensorflow jupyter
```
4. Navigate to notebook:
```bash
cd "Deep Learning Challenge"
```
5. Launch the Jupyter Notebook:
```bash
jupyter notebook AlphabetSoupCharity.ipynb
```
6. Run all cells to reproduce the preprocessing, modeling, and evaluation workflow.
---
👨💻 Developed by **Finn Brennan Arffmann**
🔗 GitHub: [@fbarffmann](https://github.com/fbarffmann)