Not every data project involves a tech company. One of the most rewarding projects I’ve worked on was helping a solid waste management company move from gut instinct to data informed decision making. The impact was tangible: trucks driving fewer miles, bins getting collected on time, and recycling rates climbing because we could actually measure what was happening.
The Starting Point
When I first looked at their operations, most decisions were based on institutional knowledge. Route planners knew which neighborhoods filled up fast on Mondays. Drivers knew which streets to avoid during school drop off hours. That knowledge was real and valuable, but it lived in people’s heads. When someone retired or called in sick, the knowledge went with them.
They had data, but it was scattered. Collection logs in one spreadsheet. Customer complaints in another. Vehicle maintenance records in a filing cabinet. Recycling tonnage reported monthly in a PDF from the processing facility. My first job was just getting all of that into a place where it could be analyzed together.
Route Optimization
The biggest operational cost in waste collection is fuel and labor, both of which are driven by how efficiently trucks move through their routes. I pulled GPS data from the fleet tracking system and overlaid it with collection completion times. The patterns were immediately visible.
Some routes had trucks doubling back through the same streets because the sequence hadn’t been updated in years. Others had imbalanced loads where one driver finished by noon and another was still working at 4 PM. A few routes included dead time where trucks sat in traffic during peak hours that could have been avoided by shifting the schedule by 30 minutes.
We restructured routes using the data, grouping nearby collection points and reordering stops to minimize backtracking. The result was a measurable reduction in total miles driven per day across the fleet. That translated directly to fuel savings and more predictable shift lengths for drivers, which was great for morale.
Collection Scheduling
The second piece was understanding fill rates. Not every bin fills at the same speed. Commercial areas generate more waste on weekdays. Residential areas spike on weekends and after holidays. Multi family housing complexes fill up faster than single family neighborhoods.
By tracking actual collection volumes over time, we could identify which areas needed more frequent pickups and which were being serviced more often than necessary. Some residential routes that ran three times a week really only needed two. Some commercial accounts that were on a fixed weekly schedule actually needed an extra midweek pickup to prevent overflow.
Matching collection frequency to actual demand reduced overflow complaints and eliminated wasted trips to bins that were only half full. It sounds simple, but when you’re running a fleet of trucks across an entire service area, those inefficiencies compound quickly.
Recycling Rate Analysis
Recycling was where the data told the most interesting story. The company reported an overall recycling rate, but when we broke it down by neighborhood, the variation was enormous. Some areas recycled over 40% of their waste. Others were below 10%.
Cross referencing recycling rates with demographic data and collection infrastructure revealed actionable patterns. Areas with lower rates often had fewer recycling bins, less convenient drop off locations, or hadn’t received educational materials in years. Some neighborhoods had high contamination rates (non recyclable items in recycling bins) that were dragging down the effective recycling yield at the processing facility.
We used this analysis to target outreach efforts and infrastructure improvements where they’d have the most impact. Placing additional recycling bins in underserved areas, running education campaigns in high contamination zones, and adjusting pickup schedules so recycling trucks arrived on days when participation was historically highest.
Why This Matters
I think about this project often because it’s a reminder that data engineering isn’t just for Silicon Valley. A waste management company might not seem like a glamorous client, but clean streets and efficient recycling are infrastructure that everyone depends on. The same analytical skills that power recommendation engines and ad targeting can make garbage collection better, and arguably that’s a more meaningful application.
The tools I used were nothing exotic. Spreadsheets for initial analysis, SQL for joining datasets, some basic visualization to communicate findings to the operations team. The value wasn’t in fancy technology. It was in asking the right questions and making the answers accessible to the people who could act on them.
Data work is at its best when it connects to something real. Trucks on roads, bins on curbs, recyclables actually getting recycled. That’s the kind of impact I want my work to have.