Open source

Built.

I cannot share really customer work, but these are some tools I have built outside project work. Free and open source.

AI skills for Open GIS

A set of AI skills for open-source GIS tools, so an AI assistant can do real geospatial work instead of guessing at it.

View project

Real-life data demos

Three quick uses of the skills above, using public data from Tartu and Tallinn.

Map of Tartu with residential buildings shaded by how many everyday services are reachable within a 15-minute walk

Tartu — 15-minute city

Each and every residential building in Tartu, scored by how many everyday services are within a 15-minute walk.

Open demo
Sources: · ETAK (Maa- ja Ruumiamet, CC-BY 4.0) · Overture Maps · OpenStreetMap (ODbL) · routing: Valhalla
Map of Tartu's Emajõgi river showing planned pedestrian and cycling bridge locations with their impact on the walking and cycling network

Tartu — bridges over the Emajõgi

Where should the next pedestrian and cycling bridge over the Emajõgi go? Each planned bridge location, scored by how much it adds to the city's walking and cycling networks.

Open demo
Sources: · Maa- ja Ruumiamet (ETAK, EHR, plans) · Keskkonnaagentuur (EELIS) · Statistikaamet · network: OpenStreetMap (ODbL), Overture Maps
Three-dimensional model of Toompea hill in Tallinn showing the castle, cathedrals and park slopes, beside the atlas controls

Tallinn — Toompea in 3D

Toompea hill and the Old Town as a source-backed 3D atlas: the city's architectural model, airborne LiDAR and aerial photography in one explorable view, with a sun and shadow study.

Open demo
Sources: · City of Tallinn (architectural model, 2020 survey) · Maa- ja Ruumiamet (2023 airborne LiDAR) · EESTIFOTO · OpenStreetMap (ODbL)

Merit Aktiva AI (and SDK)

An open Python SDK and agent interface for popular Merit Aktiva accounting software, letting AI assistants work directly with Estonian accounting data.

View project

TriFold - Discrete Global Grid System

A global grid system (DGGS) for indexing anything with a location, and to compress area data. The land/sea/population lookup libraries built on top of it run 3–30× faster than BigQuery, PostGIS and DuckDB Spatial on the same workload.

View project
Globe visualisation of the trifold hierarchical triangular grid at several levels of subdivision

Start here

Bring your own problem. Let's discuss how it can be solved.

The first step is a free 30-minute discovery call to see whether and how I can help.