Anonymous — the dedup module mirrored our Friday export chaos; still wish the pandas appendix had one more video.
Data Cleanup
CSV Rescue Lab: Cleanup Without Clicking
Normalize messy exports, detect duplicate keys, and emit a reconciliation-ready workbook.
22 hours · self-paced Self-paced with async mentor reviews BRL 540 · informational
Program narrative
Automation should feel calmer than manual clicking. You will profile messy CSVs, standardize headers, and emit annotated workbooks that highlight rows needing human eyes. The capstone mirrors a real operations desk handoff.
Inside the bundle
- Profiling notebook with histograms for string lengths
- Dedup strategies that preserve the newest row automatically
- Annotation columns that explain why a row was flagged
- Packaging the flow as a CLI your team can schedule
- Mentor session on communicating findings to stakeholders
- Template for archiving raw files before transforms
- Optional pandas track for teams already using notebooks
Outcomes we expect you to evidence
- Deliver a CLI that outputs flagged rows with reasons
- Document the cleanup assumptions your team approved
- Run the lab capstone on a masked sample from your desk
Mentor anchor
Data hygiene specialist who ships guardrails for spreadsheet-heavy teams.
Otávio Mendes
Primary reviewer for this program track.
Participant questions
No. The course targets pre-BI cleanup and reconciliation prep; dashboards stay in your existing stack.
Use masked samples inside the lab. Mentors never require production files.
We do not cover streaming pipelines; focus stays on batch CSV workflows under a few million rows.
Recent peer notes
Mentor notes on my CSV Rescue Lab CLI were blunt in a useful way — rewrote two functions same night.
Flag column wording now travels with every file I send clients.