← Back to home

Projects

Production platforms, machine learning experiments, and spatial analysis case studies — where mapping, data, and engineering meet.

Browse all repositories on GitHub →
TerraLens

TerraLens

AI-powered chatbot interface for geospatial analysis platforms. Ask questions, run analyses, and navigate maps in natural language.

LLMNLPConversational AIAPI Orchestration+1
Visit live site ↗
eSurvey Mobile App & Web Portal

eSurvey Mobile App & Web Portal

Built for Farmdar: customisable, pre-loaded crop development activity calendar with task assignment, scheduling and audit for field managers.

GDALGeoDjangoGoLangReact+2
Visit live site ↗
Erosia — Erosion Analysis Platform

Erosia — Erosion Analysis Platform

Geospatial platform for erosion prediction and sediment flow analysis, built for erosion control planning: visual erosion prediction, hydrologic analysis, and mapping for MS4 and SWPPP requirements.

GISHydrologic AnalysisSatellite ImageryErosion Control+1
Private / offline
Geospatial Project Management Platform

Geospatial Project Management Platform

Project management application for spatial data workflows: attendance tracking, project uploads, visualization tools, and delivery management for field and office teams.

ReactTypeScriptGISProject Management+1
Visit live site ↗
DuckGeoGPT

DuckGeoGPT

Query spatial databases in plain English. Gemini 2.0 translates natural language to SQL over DuckDB, with Deck.gl rendering and data export built in.

LLMGemini 2.0NL-to-SQLDeck.gl+2
Visit live site ↗
EE Health Care Facilities Map

EE Health Care Facilities Map

Interactive map of Estonian healthcare facilities against population density, supporting healthcare planning and resource allocation.

ESRI ArcGIS APIReactViteJavaScript
Visit live site ↗
Tartu Cell Towers Visualization

Tartu Cell Towers Visualization

Interactive map visualizing cell tower coverage and signal strength across Tartu using Deck.gl, React, and Mapbox.

ReactDeck.glMapboxGeoJSON+1
Visit live site ↗
Live ↗
Predicting Student Health Risk

Predicting Student Health Risk

LightGBM model classifying students into fit / at-risk / unhealthy from sleep, heart rate, BMI, calorie expenditure, step count, and exercise duration. Kaggle competition notebook, private leaderboard score 0.87498.

LightGBMGradient BoostingFeature EngineeringClassification
Visit live site ↗
Drone Path Mapping

Drone Path Mapping

Calculating and mapping efficient drone survey paths with Shapely and a greedy travelling-salesman heuristic.

GDALPythonShapely
Visit live site ↗