Portfolio
Two kinds of work: our own projects, and projects we do for clients. Client work is described without names.
Orange Lantern
One small thing a day, about two minutes long, with a painting. You say how charged you feel, one bar to three, and get one small thing sized to that. When it is done you press and hold the painting; the lantern in it lights up and the cat notices. No account, no server, no streak, nothing to type.
Native SwiftUI, local storage only, three languages with one typeface each; every painting in its own painting style. How it is made: Articles. Privacy policy · Support
Data analytics for an automotive platform
An automotive platform asked us to help them make sense of their data. They had plenty of it: listings, enquiries, dealer activity, web and app events, spread across different systems. What they did not have was one answer to simple questions, such as how many dealers were active last month. Each team kept its own spreadsheet and its own definitions.
What we did
- Data inventory. We catalogued every source: who owned it, how it was refreshed, and where it disagreed with another source. Most of the surprises surfaced here.
- Metric definitions. With the product, sales and finance leads we wrote down a single definition for each core metric, in plain language and in SQL, and parked the ones nobody could agree on.
- Data pipeline. Scheduled jobs that pull from the source systems into a warehouse, with tests that fail loudly when a feed changes shape.
- Dashboards. A small set of boards built on the agreed metrics, one per team, answering the questions that team actually asks.
- Handover. Documentation, a runbook, and pairing sessions so the client's engineers could run the pipeline without us.
How we worked
By the week, remotely, inside the client's own tools: their repository, their chat, their ticket board. Dako paired with their analyst and backend engineers. Each week ended with something usable and a short review call to decide what came next.
What we delivered
A documented inventory of sources, a shared metric dictionary, the pipeline code with its tests, the dashboards, and a runbook for keeping all of it alive.
What their team took over
By the end, their engineers were adding new sources themselves and their analyst owned the metric dictionary and dashboards. We stayed on as an occasional second opinion, then stepped back.
Working with Dako on analytics
We take on data projects by the week, usually remote. The shape is always the same: inventory, definitions, pipeline and dashboards, then a handover that leaves your team in charge. If your numbers change depending on who you ask, write a few lines to hello@dako.world.