public-source families
including GUGiK, Wody Polskie, PIG-PIB, GDOŚ, OSM and related public services
KAIM / SYSTEM 02 · ACTIVE INDEPENDENT PRODUCT
Okolica Działki
A parcel-level property intelligence system for evaluating residential land in Poland. A selected point is resolved to the actual cadastral parcel, then mobility, planning, terrain, environmental and risk evidence are normalized into an explainable decision model.
KAIM / BUSINESS VALUE
Okolica Działki compresses property research into one inspectable flow. Instead of opening separate maps, registries, planning portals, terrain and environmental data, and routing tools, the user starts from a real cadastral parcel and gets evidence assembled around it.
decision domains
everyday utility, mobility, planning, environment, terrain and risk in one model
parcels in deterministic comparison
the same decision model can be applied to several real purchase alternatives
WHAT CHANGES FOR THE DECISION
The main cost of parcel research is rarely access to data itself. It is finding the right evidence, attaching it to the correct parcel and deciding what missing or conflicting evidence actually means for the purchase decision.
The analysis starts from a real cadastral parcel, not a map pin.
A point is resolved to a cadastral parcel through ULDK. Parcel geometry becomes the shared context for planning, terrain, risk, infrastructure, mobility and reporting.
Missing evidence stays missing.
Evidence keeps provenance, availability and limitations explicit. Uncertainty is not converted into a silent positive result, which matters when the underlying purchase can cost hundreds of thousands of złoty.
The same parcel can be a good or bad decision under different life scenarios.
Lifestyle profiles, three mobility modes, multi-parcel comparison and geometric house-footprint testing move the product from a generic score toward actual decision support.
KAIM / TRUST
Scope of responsibility, without inference.
- STATUS
- ACTIVE INDEPENDENT PRODUCT
- ROLE
- Product architect / Full-stack engineer
- PERIOD
- Aug 2026–present
- OWNERSHIP
- Product concept, geospatial architecture, source integrations, parcel intelligence model, scoring, routing, comparison engine, terrain analysis, House Fit, evidence model, frontend, edge/API runtime, automated tests, quality gates and deployment workflow.
- NOT OWNED
- Underlying public datasets, cadastral records, planning acts, flood models, OpenStreetMap data and third-party routing infrastructure.
- LIMITATIONS
- The application aggregates and interprets public evidence. It is not a legal opinion, cadastral survey, geotechnical report, utility-connection guarantee or building-permit decision.
- VERIFICATION
- Automated unit and regression tests, source-contract monitoring, numeric sanity guards, runtime healthcheck and deterministic evidence/model inspection.
KAIM / EVIDENCE
Every claim carries an epistemic status.
Numbers are not treated as equivalent. COUNTED means counted in an artifact or repository, MEASURED means measured in a specific workflow, ESTIMATED is approximate, and DOCUMENTED marks a qualitative workflow or architecture contract.
DOCUMENTED 6 decision domains
ProvenanceProduct domain model · Daily life / Mobility / Planning / Environment / Terrain / Risk
DOCUMENTED 10+ public source families
ProvenanceRuntime source adapters and Evidence Stack
ScopeIncludes GUGiK, Wody Polskie, PIG-PIB, GDOŚ, OSM and related public services; availability varies by parcel and source.
DOCUMENTED 60 automated regression tests
ProvenanceProject quality state around generation 6.2.x
ScopeDocumented development-state count; re-audit before any future publication update.
DOCUMENTED 3 plots in deterministic comparison
ProvenanceCompare engine product contract
DOCUMENTED 3 mobility modes
ProvenanceCar / walking / bicycle
DOCUMENTED 8 due-diligence report sections
ProvenanceProperty Due Diligence Report structure
KAIM / SYSTEM MAP
The architecture in one pass.
- 01Map / search → ULDK → cadastral identity and polygon
- 02OSM / Geofabrik → POI normalization → routing → lifestyle-aware utility
- 03Planning / NMT / land use / utilities / flood / landslide evidence
- 04Six domains → overall score → strengths / weaknesses → confidence
- 05One parcel × multiple lifestyle scenarios
- 06Parcel polygon × house footprint × terrain
- 07Compare / inspector / report / GeoJSON / source contracts
System flows / 05
How a map click becomes parcel-level evidence.
A simplified cut through the live product: a location resolves to the actual cadastral parcel, public GIS and mobility sources are normalized behind a controlled API, scoring stays deterministic, and completeness plus provenance flow into decision tools and export.
An address or map click starts a parcel-first analysis together with the user's profile context.
- address / map
- profile
- local state
The point resolves to the identifier and geometry of the actual cadastral parcel; a missing parcel does not collapse the whole analysis.
- parcel ID
- polygon
- fail-soft
Terrain, planning, land use, environment, risks, utilities and access arrive as independent evidence branches.
- terrain
- planning
- risks
The mobility path joins nearby places with real or explicitly estimated car, walking and bicycle travel time.
- POI
- route matrix
- fallbacks
Profile, POI, routing and parcel intelligence combine into six domains, an overall score and confidence without an LLM generating the decision.
- 6 domains
- 0–100
- confidence
Every signal keeps its source, availability state, timestamp and completeness; unknown is never silently coerced to zero.
- provenance
- available / partial
- unknown explicit
Compare, Scenario Lab, House Fit, Terrain, Access and History work from the same parcel facts rather than separate narratives.
- compare
- scenario
- house fit
Decision support ends in a property brief and GeoJSON while retaining sources and methodological limitations.
- property brief
- GeoJSON
- method limits
Flow point Flow point
KAIM / ENGINEERING
What mattered technically.
Property decisions depend on data that lives in completely different systems.
A cadastral parcel, its terrain, planning context, flood exposure, utilities and everyday accessibility usually have to be checked independently. Okolica Działki turns these disconnected sources into one parcel-centered decision model.
The cadastral polygon is the domain object.
A map interaction resolves a real parcel through ULDK. From there, asynchronous adapters query mobility, planning, terrain, environmental and risk sources. Each observation is normalized into an evidence model that carries provenance, availability and retrieval context.
The score is only the surface; evidence remains inspectable.
Users can drill into domain scores, compare three parcels, replay different lifestyle scenarios, inspect terrain and test a house footprint geometrically. Missing evidence remains unknown instead of becoming a silent positive result, while source-contract tests protect integrations against schema and numeric regressions.
Source / influencePublic evidence is normalized with explicit availability, provenance and limitations; missing data is never treated as proof of a positive outcome.