FormFix AI is a serverless data cleaning and sentiment analysis tool that uses Python, OpenAI, and AWS to transform inconsistent form responses into structured, readable formats.

• ✅ Standardizes inconsistent fields (e.g. "maybeee", "IDK", "nah fam") into Yes, No, Maybe, or Unknown
• 🧠 Uses OpenAI for intelligent classification and analysis
• ☁️ Serverless deployment with AWS Lambda and Docker
• 📂 Fetches and stores data via AWS S3
• 🔁 One-command execution from terminal

1. Upload your CSV to an S3 bucket
2. Trigger the Lambda function from your terminal using a bash script
3. The function:
   • Retrieves the file from S3
   • Extracts the necessary column using pandas
   • Sends it to OpenAI for context-aware analysis
   • Writes the cleaned output back to S3 as a new CSV

• Python 3.11
• OpenAI API (gpt-3.5-turbo)
• AWS Lambda (Dockerized)
• AWS S3
• Libraries: boto3, pandas, openai, python-dotenv

formfix-ai/
├── .gitignore
│── dirty_data_sample.csv        # Lambda function handler
│── form_fix_ai.py          # Utility functions for S3 and OpenAI
├── Dockerfile            # AWS Lambda base image
├── requirements.txt      # Python dependencies
├── .env                  # Environment variables (local only)
├── local_lambda_deploy.sh                # Bash script to build & deploy
├── README.md             # Project documentation

git clone https://github.com/yourusername/formfix-ai.git
cd formfix-ai

Create a .env file and add:

OPENAI_API_KEY=your_openai_api_key
AWS_ACCESS_KEY_ID=your_aws_key
AWS_SECRET_ACCESS_KEY=your_aws_secret
S3_BUCKET_NAME=your_bucket_name
S3_INPUT_FILE=input.csv
S3_OUTPUT_FILE=cleaned.csv
TARGET_COLUMN_NAME=your_column_to_clean

bash local_lambda_deploy.sh

You’ve collected survey responses with inconsistent answers in a column like "Availability" — responses like:

"yes", "yeah", "nah", "no", "idk", "maybeee"

FormFix AI will convert those into clean values:

"Yes", "No", "Maybe", "Unknown"

Feel free to fork this repo, modify the prompts, and adapt the pipeline for your own use cases. Want to add a GUI or streamlit frontend? Go for it.

If you share your remix or build publicly, tag with #AIWithFafa so we can connect and amplify! 🚀

GNU GPL License

Fafa Modey — Python Engineer, AI Builder, and Digital Rights Advocate

Follow my journey on Instagram: @fafacodes

Project page: [Link in bio]