Self-Service ML Data Platform

Django-based platform for uploading, cleaning, and visualizing multi-format datasets with real-time feedback.

The problem

Analysts needed a self-service way to upload heterogeneous data files (CSV, XLSX, Parquet, JSON) and get automated cleaning and visualization without engineering support.

Impact

Data preparation time: -10%
Formats supported: CSV, XLSX, Parquet, JSON

Approach

1
Built authenticated file upload and processing flows with automated cleaning, filtering, and transformation.
2
Delivered real-time visualization with Matplotlib, Seaborn, and Plotly.
3
Adopted a microservice architecture with Docker and Celery/Redis for background processing.

Stack

Django
Docker
PostgreSQL
MySQL
Plotly
Celery
Redis