DataLoom

DataLoom is a browser-based data wrangling workspace that enables users to upload tabular datasets, apply reversible transformations, and manage data visually without writing code, supported by a checkpoint and revert system. Currently at the MVP stage, DataLoom supports only CSV files and lacks key capabilities such as data profiling, visualization, dataset merging, and efficient column selection. This project expands DataLoom into a full data preparation platform by introducing multi-format ingestion, automated data profiling, dataset joins and concatenation, formula-based transformations, reusable pipelines, interactive visualizations, and an automated data quality engine with scoring and one-click fixes. It also adds multi-format export and downloadable reports. Key deliverables include multi-format ingestion and export, a profiling API with a reusable column selector, join and merge operations, a visualization panel, a data quality engine, a formula editor, reusable pipelines, downloadable reports, a full frontend TypeScript migration, a refactored backend, comprehensive testing, and a redesigned scalable UI. The outcome is a robust, end-to-end data preparation tool that supports the complete workflow from raw data ingestion to cleaned, analyzed, and export-ready datasets.