Roadmap: data prep and portability backlog

[daniel-thom]

Summary

Track the initial data prep and portability roadmap for datasight.

This backlog focuses on helping users move from raw files to analysis-ready, portable, and shareable datasight projects.

Proposed sequence

• Add managed import modes for CSV and Parquet ingestion #46 Add managed import modes for CSV and Parquet ingestion
• Add export bundles with SQL, Python, HTML, and metadata #49 Add export bundles with SQL, Python, HTML, and metadata
• Add session export and import for shareable datasight workspaces #50 Add session export and import for shareable datasight workspaces
• Detect untidy datasets and guide users toward tidy reshaping #51 Detect untidy datasets and guide users toward tidy reshaping
• Add a prep recipe engine for inspectable data cleaning workflows #47 Add a prep recipe engine for inspectable data cleaning workflows
• Expand keyboard shortcuts and command palette coverage #48 Expand keyboard shortcuts and command palette coverage

Why this order

1. Import modes are immediately useful and improve performance/portability with relatively low implementation risk.
2. Export bundles and session export/import make analyses easier to share, archive, and reproduce.
3. Untidy-data detection opens the door to guided preparation workflows.
4. A prep recipe engine provides the inspectable execution model needed for AI-assisted reshaping and cleaning.
5. Shortcut expansion is independent polish that can ship whenever convenient.

Outcome

Completing this roadmap should give datasight a more coherent flow from:

• raw CSV/Parquet inputs,
• to managed and portable project state,
• to reproducible outputs,
• to guided data preparation for messy real-world datasets.