Knowing what every cleaning job really takes.
Fresh Focus cleans trash chutes, compactor rooms, carpets and more for apartment and commercial buildings from New York to Washington, D.C. I built a tool that shows how much crew time each job actually takes.
- Role
- Consultant
- Years
- 2026
- Based in
- Barrington, NJ
The problem
To price jobs well, Fresh Focus needs to know how long each crew spends at every job, including the driving to get there. Working that out meant going through GPS history by hand, one van at a time.
What I built
- Timesheet pipelineA Python tool that combines scheduled visits, GPS trips from each van and drive times from Google Maps into one timesheet row per visit.
- On-site timeMatches each van's stops to job sites, so working time is counted as time on site rather than driving.
- Drive timeSplits each day's real driving time fairly across its jobs.
- AI for the edge casesWhen the GPS data is messy, the tool flags the visit and Claude Code reads the track to fill in the gap, recording its reasoning next to the numbers.
Outcome
What used to be manual work is now a report with on-site and drive hours for every visit, already calculated. Only the occasional unclear job needs a closer look. I also wrote plain-language docs so the owners can see how every number is worked out.
Mike was easy to work with from the start. He took a process we were doing by hand, van by van, and turned it into a report that's ready when we need it. We finally know what our jobs really take.