Methodology
All displayed NSW and Queensland crime figures are official data. Queensland pages use the most detailed available source: a crime-map postcode export where available, otherwise a clearly labelled police division baseline with LGA context. Wider-area figures are not individual suburb counts; they are not sample data.
Datasets are discovered each refresh through the Queensland Government Open Data Portal CKAN package_search and package_show APIs, restricted to organisation “Police”. Resource URLs are resolved from current package metadata rather than hard-coded S3 links.
Last successful baseline refresh: 10/11/2026, 9:33:50 pm AEST. Monthly schedule: 15th of each month at 03:00 Australia/Brisbane. Next due: 15/11/2026, 3:00:00 am AEST.
The API service checks the database-backed due date hourly and catches up after downtime; a database advisory lock prevents concurrent imports. Exact unattended scheduling requires the service to remain running, or the supplied one-shot refresh command to run in scheduled hosting. An unpublished/sleeping development workspace cannot guarantee a clock-time run. No schema migrations run in the refresh job.
QPS Divisions boundaries come from the same portal. ABS 2021 SAL/POA polygon centroids are assigned to containing QPS divisions; multipart/coastal features whose centroids fall outside use largest actual boundary overlap in Australian Albers. QLD LGA mappings use the largest shared land area in ABS mesh-block allocations. Boundaries can span multiple areas; the primary assignment does not imply exact containment. Boundary and Census source hashes and all mapping exceptions are retained; no nearest-division guesses are silently substituted.
The area CSVs contain hierarchical parent totals and their descendants. Only explicitly reviewed, non-overlapping parent offences are mapped into nine categories. Child columns are not added again. Every row is reconciled against the three published offence-group subtotals; the complete 92-column mapping audit is retained with the import. Unknown columns, suppressed counts or broken reconciliation stop a refresh without replacing good data.
QPS publishes division/LGA rates per 100,000 persons. We sum the published monthly rates for the displayed twelve months and divide by 100 to display per 1,000, using the source's area population basis. Where a population label is shown, it is inferred from the published counts/rates and rounded, not an ABS suburb Census denominator. Division/LGA counts are never divided by the population of a smaller suburb or postcode. The separate LGA panel has its own clearly stated period; its counts must not be added to the division or postcode figures.
The Queensland Police Service publishes offence counts by location and month as open data under a Creative Commons Attribution (CC BY) licence.
Queensland Police Service maps and statistics. The actual Online Crime Map XLSX/CSV export has three columns: Offence Type, Offence Date and Suburb. Despite its label, Suburb contains a postcode. One row is one offence; there are no offence IDs, coordinates, times of day or street addresses. We aggregate by postcode, category and date, then month, without inventing incident map points.
All ABS QLD suburb and postal-area locations have catalogue entries. Suburbs with a covered postal area use that postcode's export, explicitly labelled as postcode—not suburb—counts. Other suburbs retain their own URLs with division fallback. Lists use SAL/POA mesh-block correspondence; intersecting postcodes are ranked by shared land area. Export rates use the national ABS 2021 POA G01 usual-resident population, not a sum of rounded suburb populations. POAs approximate postal geography, not an address directory.
Each export records its SHA-256 hash, exact inclusive date coverage, postcodes, row counts and offence totals. Reimporting identical bytes is idempotent. There are no stable offence identifiers: indistinguishable rows within one file are preserved as separate offences. For overlapping, complete, unfiltered postcode/date exports, the newest imported snapshot replaces older counts for the overlapping days, including zero-count days. This avoids double-counting overlapping snapshots without falsely deleting repeated offences. It is not individual-incident identity matching; incomplete or filtered exports must not be represented as complete postcode coverage.
Our first workbook contains 28,087 rows from 10 November 2025 through 10 November 2026 inclusive. The accompanying PDF labels its end date 9 November, but the workbook has one offence on 10 November; all workbook rows are retained. The current day can be incomplete. First/last calendar months are partial; charts label available coverage, not an invented five-year history. QLD rates divide counts for the exact displayed period by Census population and are not annualised. Unavailable prior history is shown as unavailable, not zero. Border postcode 2406 is retained internally but only gets a public QLD page at 100 or more imported offences.
BOCSAR publishes recorded criminal incidents by suburb and month for New South Wales, also under CC BY.
BOCSAR criminal incident datasets. The NSW importer reads the monthly SuburbData.zip dataset and maps its 62 offence types into nine unified categories. Unknown offences or missing counts stop an import rather than being silently counted as zero.
Names and polygon centroids come from ABS Suburbs and Localities (SAL) 2021. Resident population is the ABS 2021 Census General Community Profile, G01 total persons (Tot_P_P), by usual residence. Census counts are privacy-perturbed. These are fixed 2021 denominators, not estimates of today's population; growth since 2021 can affect rates.
Postcodes are ABS Postal Area (POA) 2021 approximations, not an Australia Post address directory. SALs are linked through shared ABS mesh blocks to postal areas and 2021 LGAs. Where there are several, the primary label is the one with the largest shared land area; all linked postal areas and LGAs are retained for search. Centroid distances are straight-line statistical comparisons, not travel distances. ABS SAL boundaries may differ from police locality boundaries, and name matching does not guarantee identical geographic boundaries.
ABS 2021 allocation files and 2021 NSW SAL Census DataPack, © Commonwealth of Australia, Australian Bureau of Statistics, CC BY 4.0. Full source provenance and unmatched-name reports are retained with the import tools.
Break-ins, Vehicle theft, Theft, Assault, Robbery, Property damage, Drug offences, Fraud and Other. Sexual offences, homicide, intimidation and justice-related offences are included in Other, not in Assault. Theft from a vehicle is Theft; theft of a vehicle is Vehicle theft. These groupings simplify presentation and do not make QLD and NSW classifications directly equivalent.
Source: Queensland Police Service / NSW Bureau of Crime Statistics and Research, CC BY licence. Figures are recorded offences and may not reflect all crime.