encode This

So I built something new. I called it... encode This

As a data engineer, I spend most of my time wrangling data pipelines—but on free days, I like to build tools that solve real problems. With the power of AI building solutions becomes more possible.

My latest side project is encode This: a web app designed to help researchers and analysts upload survey responses in Excel and automatically encode the data for statistical analysis. No more manual mapping or tedious formatting—just upload, encode, and go.

This project was born out of a simple need I noticed, and I built it from with help of some clever AI promptiong, only to make data prep faster and easier. It’s a reminder that with the right mindset, we can always learn, adapt, and build solutions on the go.

The tech stack used is as follows:

Frontend AI Design

• Gemini

Backend AI Design

• Claude

Frontend

• Static HTML5 with vanilla JavaScript (no framework, no build step)
• Tailwind CSS via CDN for styling
• Inter Google Font
• Served by the backend at /app using StaticFiles
• File upload via FormData; responsive layout with collapsible panels

Backend

• Python 3.11.x
• FastAPI (API framework) + Uvicorn (ASGI server)
• Pydantic v2 (request/response models)
• Endpoints include: /health, /schema, /encode/preview, /encode/process, /encode/dictionary

Data and encoding libraries

• pandas (data loading/processing)
• numpy (numerics)
• scikit-learn (e.g., OneHotEncoder and related utilities)
• rapidfuzz (fuzzy matching for canonicalization)
• openpyxl (Excel read/write, multi-sheet support)

As the weeks go, we will see what else can be built.

#DataEngineering #WebDevelopment #SideProject #EncodeThis #ProblemSolving #StatisticalAnalysis #TechForGood