Alphabet Soup Charity Analysis

Description

This project uses a deep learning neural network to classify whether a charity funded by Alphabet Soup will be successful. Leveraging TensorFlow, Pandas, and scikit-learn, the model performs binary classification based on key categorical and numeric indicators.

Summary

Part 1: Data Preprocessing

The dataset was cleaned and prepped using standard data science practices.

Key Deliverables:

• Load charity_data.csv into a Pandas DataFrame from the provided cloud URL.
• Drop unnecessary columns: EIN and NAME.
• Identify the target (IS_SUCCESSFUL) and features (all remaining relevant columns).
• Examine the number of unique values per column.
• Consolidate rare categorical values under an "Other" category.
• Encode categorical features using pd.get_dummies().
• Split data into training and testing sets using train_test_split.
• Scale feature data using StandardScaler() from scikit-learn.

Part 2: Neural Network Model

A deep learning model was built using TensorFlow/Keras to classify successful vs. unsuccessful organizations.

Key Deliverables:

• Define input dimensions based on the number of features.
• Build a sequential model with:
  • One or two hidden layers (ReLU activation).
  • One output layer (sigmoid activation).
• Compile the model using binary crossentropy and the Adam optimizer.
• Train the model on training data, saving weights every 5 epochs using a callback.
• Evaluate the model against test data for loss and accuracy.
• Save final model as AlphabetSoupCharity.h5.

Part 3: Model Optimization

Multiple attempts were made to improve accuracy beyond the 75% target.

Key Deliverables:

• Adjusted model architecture:
  • Added/removed layers and neurons.
  • Modified activation functions.
  • Tuned the number of training epochs.
• Adjusted data preprocessing:
  • Modified how rare values were grouped.
  • Reconsidered which features to drop.
• Final model saved as AlphabetSoupCharity_Optimization.h5.

Note: A formal written report was not completed for this project. All findings are documented within the Jupyter notebooks.

Requirements

Tools and Libraries

• Python
• Google Colab / Jupyter Notebook
• Pandas
• scikit-learn
• TensorFlow / Keras
• HDF5

pip install pandas scikit-learn tensorflow h5py

Setup

Conda Environment

To recreate the working environment for this project:

bash conda create -n alphabet_env python=3.9 conda activate alphabet_env pip install pandas scikit-learn tensorflow h5py

Contact Information

For questions or additional information, reach out to me at:

• Email: ilir.hajdari111@gmail.com