What these projects actually look like
Four engagements that show the shape of the work: what was going wrong, what I built, and what changed once it was live. The reporting views alongside them are mockups, drawn to show the kind of thing I build rather than any one client's numbers.
Four days of month end, gone
The situation
A finance team of five spent the first four working days of every month rebuilding the same board pack by hand. Figures were exported from the ledger, reconciled in Excel, then pasted into slides. When a number was questioned in the meeting, nobody could answer on the spot, because the trail ran back through six workbooks and one person's memory.
What I did
Modelled the general ledger into a star schema with a proper date table, so periods, comparatives, and year to date all behaved consistently instead of being recalculated by hand each time. Automated an overnight refresh from the finance system. Rebuilt the pack in Power BI, keeping the commentary boxes the board actually reads, and added drill through to the underlying transactions on every figure.
What changed
The pack is ready on the first working day, and it says the same thing every month. When a figure gets queried, someone clicks into it while the meeting is still running.
Illustrative view, not client data
One screen instead of three systems
The situation
Dispatchers at a field service business tracked jobs across a scheduling tool, a spreadsheet of engineer availability, and a shared mailbox. The status of a job depended on who you asked. Customers were being told an engineer was on the way when the job had already been reassigned that morning.
What I did
Built Fabric pipelines to bring the scheduling system, the staffing feed, and the job notes into one lakehouse, refreshing through the day rather than overnight. Designed a single operational view showing jobs by status, engineer, and region, with exceptions surfaced at the top instead of buried somewhere in a table of four hundred rows.
What changed
Dispatch works from one screen. The morning stand up and the afternoon review look at the same numbers, which turned out to matter more than anyone expected.
Illustrative view, not client data
A team that stopped waiting on one person
The situation
In a company of sixty, every report request queued behind a single analyst who properly understood the data. She was three weeks behind, and most of what she was working through were small changes that mattered to someone for about a day. Nobody else felt confident going near the numbers.
What I did
Built two governed datasets covering sales and operations, with the measures defined once and row level security applied so people only see what they should. Ran three workshops teaching the wider team to build their own reports on top, and wrote a short guide covering the handful of things people reliably get wrong with dates and filters.
What changed
Most requests never reach her now. The ones that do are the genuinely hard problems, which is a considerably better use of the only person who understands the data end to end.
Illustrative model, not client data
Off the ageing warehouse, without a big bang
The situation
Reports were slow, the warehouse was expensive to keep running, and replacing it had been quoted as a year long programme that nobody wanted to be the one to start. So it kept not starting, and the running costs kept arriving.
What I did
Migrated in slices rather than all at once. The highest value subject area moved into a Fabric lakehouse first, then ran in parallel with the old platform for a month so we could prove the numbers matched before anyone relied on it. Then the next slice, the same way. Each one was small enough to reverse in an afternoon if it went wrong.
What changed
Reporting got faster in stages instead of all at once at the end of a long programme. The old platform was switched off once it was empty, rather than on a date somebody had picked a year earlier.
Illustrative plan, not client data
A note on these examples
Databert is a young consultancy, and I would rather show you the shape of the work honestly than dress up someone else's case studies as my own. The four above are representative engagements drawn from the kind of problems I take on, and the reporting views are mockups built for this page. As real projects complete and clients are happy to be named, they will replace these.
If you would like to talk to someone who has worked with me, ask on our first call and I will put you in touch.