Hatched areas: distance-weighted estimates within 2 km of mapped reports. Missing reports and history are not reconstructed.
Drag or zoom to move all four maps. Shared scale.
Green: low · red: high · gray: unknownClick a count to filter the map.
Empty hatching = no loaded reports that month, not zero crime. Click a count to inspect that month and category on the map. Agencies have different date bases and incomplete log windows.
Map windows end today (UTC). Sheriff uses log dates; other sources use occurrence dates where available. All time includes only loaded records, not complete crime history.
Mapped records combine Santa Cruz Police, County Sheriff and Watsonville logs with UCSC, Scotts Valley and Capitola public crime-map cases. Agency date windows differ. Capitola adds a recent public crime-map sample; a Sheriff report in a town does not represent all local crime. Community filters use reported localities, which can be postal areas rather than administrative boundaries.
SCPD logs contain reports approved in the previous 24 hours and remain online for roughly 30 days. Sheriff PDFs use their own Date/Time field; Watsonville scans require OCR and verification against the originals. The coverage strip shows the loaded sample for each selected agency. A recent publication can describe an older event. This atlas preserves occurrence, reported, and publication dates separately.
UCSC, Scotts Valley and Capitola publisher coordinates are generalized to 250 m cells; out-of-area coordinates remain unmapped. PDF police locations are matched to the average of official address points on the published block, using county addresses for Sheriff and Watsonville records. Exact street numbers are reduced to blocks. This is an approximation, not an exact incident location. Intersections and unmatched blocks stay in the list without an invented pin.
Reports and arrests are not convictions. Names, ages, and narrative columns are excluded. Code enforcement target areas are administrative boundaries, not individual violations. Missing sources never become zero incidents. Geographic heatmaps sum compact, distance-weighted report kernels around approximate locations. The radius is a display choice, not a measured risk boundary. Green indicates lower observed report intensity; it does not mean safe. Unobserved areas remain neutral gray. Smoothing is visual context, not predicted crime or exact locations. Severity mode multiplies each kernel by an editable category weight (homicide 10, sex offenses 8, robbery 6, assault 5, weapons 4, threats/burglary/vehicle theft 3, others 1). These are editorial weights, not an official harm index. Report-count mode gives every report weight 1; unmapped reports are excluded. Comparison maps share extent, smoothing and color scale. The optional nearby-gap layer extends a distance-weighted estimate up to 2 km from mapped reports. It uses a broader biweight kernel, scaled by the squared radius ratio to preserve total kernel mass. Only gaps in the main heat surface are drawn, with diagonal hatching. This is spatial smoothing, not a reconstruction of missing reports, reporting coverage, or history. Areas beyond nearby observations remain unknown. Monthly heatmaps count reports, including unmapped ones. Blank months mean no loaded reports, not zero crime. Annual DOJ totals are separate offense aggregates; they are never assigned to map locations.
Independent portfolio project. Not affiliated with the City of Santa Cruz. Basemap © OpenStreetMap contributors. Data remain subject to publisher terms and revisions.