ExtracaoRDA
A desktop app that turns PADIS reports into validated spreadsheets and removes manual transcription from the workflow.
A solo project conceived and built during my MCTI internship, outside the team scope.
Python · PyQt · PyInstaller · pytest
- 0 manual transcription
- Cross-validation
Context
During my internship at MCTI, analysts had to manually transcribe data from PADIS Annual Reports into spreadsheets. ExtracaoRDA was an individual initiative, created outside the NOVO SIGPLANI team scope.
The problem
The manual workflow consumed time and allowed inconsistencies that were difficult to spot. A useful automation had to extract values from PDFs, structure the spreadsheet and verify the result against other sources.
My contribution
I conceived and developed the Python and PyQt desktop application on my own. I implemented extraction, Excel generation, cross-validation, pytest tests, PyInstaller packaging and continuous integration with GitHub Actions.
Result
Manual transcription left the analysts’ workflow. The validation layer flags inconsistencies for review, keeping the final decision visible instead of treating automated extraction as infallible.
The repository is private because the work was developed in the context of the internship.