Messy data in. Clean reports out.
A working simulation of the data pipelines I build. Send in spreadsheets, exports and bank files, switch the cleaning rules on and off, and watch what reaches the report.
The pipeline
Files travel left to right. Click a file to inspect it below.
Waiting for the next file
Inspector
Waiting for a file
Swipe the table sideways to see every column.
Cleaning rules
All five rules are on
Latest report
Waiting for the first file
The first report appears in a few seconds.
Illustrative model: manual time assumes roughly 20 to 40 minutes of copying, fixing and checking per file.
Activity
Gold entries were handled by a rule. Red entries need attention.
Built from scratch for this site with TypeScript and no libraries. The company, files, names and numbers are invented, and the time-saving figures come from a simple illustrative model, not a measurement.
Every file checked the same way
The rules never get tired, rushed or creative. A price list is read on a Friday afternoon exactly as it is on a Monday morning.
Doubtful data is held, not guessed at
Records with missing or suspicious values are set aside for review. Your reports stay trustworthy because nothing is quietly invented.
Failures that recover or tell you
Connections drop. The pipeline retries with sensible waits, recovers on its own when it can, and raises a clear alert when it cannot.
Tired of fixing spreadsheets by hand?
Book a private 30-minute call and describe the reports your team rebuilds every week. You will leave knowing what could run by itself and what it is worth.