A practical workflow for lenders, insurers and enterprise real-estate risk teams

Bring geographic exposure into an enterprise property review

Reviewing a large property portfolio one address at a time is costly. Before commissioning property-specific hazard reports, enterprise teams often need a consistent way to see where geographic exposure is concentrated and which clusters deserve attention first.

An H3 screening layer can organise that first pass. It divides a selected local government area into consistently identified cells, adds flood and bushfire land-overlap observations, and gives a portfolio team stable identifiers for an offline join to its own aggregated property counts. The result is a traceable review queue.

This is an area-screening workflow. It does not determine whether a parcel is affected, decide whether a property is insurable, establish lending eligibility or support an inference about a borrower, owner or occupant.

Start with one common geographic frame

The workflow resolves Lismore City Council, NSW to NSW236, then retrieves its resolution-7 H3 concordance. The returned concordance contains 178 cells. Applying the same min_n=20 support threshold and strict resolution to both exposure surfaces produces 142 cells with complete flood and bushfire observations.

The public map is a bounded excerpt of the 30 cells with the highest shared returned support. The selection rule uses support only; it does not use the exposure values. Every displayed polygon comes from the H3 boundary operation and stays attached to its exact cell identifier.

That common frame matters because an H3 cell can cross familiar suburb or administrative boundaries. A cell result remains a cell result; it should not be relabelled as a parcel, suburb, household or borrower observation.

Read both measures as land overlap

The live endpoint contracts define pct-flood as the percentage of each H3 cell overlapping flood-prone land and pct-bushfire as the percentage overlapping bushfire-prone land. The API returns decimal fractions, so the demo multiplies each value by 100 once for display.

Across the 142 returned supported cells, the median flood-prone land overlap is 17.0% and the median bushfire-prone land overlap is 71.5%. These are medians of cell-level area-overlap observations in this captured response. They are not LGA parcel-exposure rates, event probabilities or severity measures.

Keep observation, interpretation and action separate

Observation: cell 87bea9611ffffff has 91.0% flood-prone land overlap and 84.0% bushfire-prone land overlap. Both exceed the captured cell medians.

Interpretation: under the stated queue rule, the cell deserves earlier geographic review because both area-level observations sit above their comparison medians. That ranking describes the selected cell and rule; it does not establish exposure for any property inside the cell.

Suggested action: route the relevant internal property record to the organisation’s established parcel-level hazard, valuation, insurance or resilience review. Keep the H3 observations as provenance for why the record entered the queue, while recording the separate evidence and decision produced by detailed review.

Reuse the workflow without rebuilding the map

Fetch the two hazard surfaces once for the chosen LGA and H3 resolution, cache the reviewed snapshot under the organisation’s retention policy, and reuse the exact cell identifiers when joining an offline portfolio aggregation. Keep the LGA identifier, resolution, support threshold, capture time, unit conversion, selection rule and queue rule with every saved review.

The accepted production workflow uses five distinct HtAG operations across 34 requests and returns 493 rows. Thirty requests retrieve the selected cell boundaries. Distinct operations, requests, rows and billing are separate measures; current pricing should be checked rather than inferred from route count. Reader interaction loads map tiles but makes zero reader-time HtAG API calls.

Implementation appendix

For the H3 cell format, all supported resolutions and the geography-mapping contracts used to assemble this review, read our H3 format, endpoints and geography-mapping guide.

The HtAG API documentation covers the five operations:

1. GET /reference/lga — resolve the reviewed LGA identifier.
2. GET /reference/concordance/lga-to-h3 — retrieve H3 cells for the LGA and resolution.
3. GET /spatial/h3/socio-env/pct-flood — retrieve cell-level flood-prone land overlap.
4. GET /spatial/h3/socio-env/pct-bushfire — retrieve cell-level bushfire-prone land overlap.
5. GET /spatial/h3/cells/{h3_index}/boundary — retrieve fixed polygon geometry for the selected map excerpt.