02
Dataset analysis

Wrangler

Wrangler profiles CSV, Excel, Parquet, or Google Sheets data, applies selected cleaning and aggregation steps, and produces a React and Plotly dashboard with retained session artifacts.

Live
  • Python
  • FastAPI
  • React
  • Plotly
  • Data quality
Visit product ↗
01 / Context
Problem

Before analysts can trust a dashboard, they must find missing values, duplicate rows, outlier candidates, and unsuitable aggregations in unfamiliar data.

My work

Product direction, data-quality pipeline, FastAPI orchestration, dashboard specification, and the React interface used to review and refine the result.

02 / Architecture
01Dataset upload
02Null + duplicate + outlier profiling
03Cleaning + aggregation plan
04FastAPI generation
05React + Plotly dashboard
06Saved session artifacts
03 / Key decisions
01

Quality findings and transformation choices remain visible beside the charts, so the dashboard can be reviewed against the source data.

02

The visualization plan maps selected metrics to chart types, filters, aggregations, and layout sections before rendering the final dashboard.

03

The live product includes upload, review, results, and saved-session flows, with bundled datasets that require no private information.

04 / Screenshots