The story in one minute

Family violence reports to Victoria Police rose to a record 106,430 incidents in 2024-25. The headline number hides a split. Across metropolitan Melbourne the rate of incidents has barely moved in five years. Across regional Victoria it has climbed by about 18%, and in some towns by more than half.

This report follows that gap from the state level down to individual towns, and then asks a harder question: as demand rose in the regions, did specialist help keep pace?

Reading tip: every chart is interactive. Hover for details, click legend items to hide or show groups, and drag to zoom (double-click to reset).

If this report raises something for you: 1800RESPECT (1800 737 732) is available 24/7. In an emergency call 000.

1. Two Victorias

Chart 1. Population-weighted rate of police-recorded family violence incidents. Metro = eight Melbourne DFFH regions; Regional = the nine others. Source: Crime Statistics Agency (2025), Tables 1-2.

A rate matters more than a count here. Melbourne’s number of incidents grew by about 8%, but its population grew by about 9%, so the rate is flat. Regional Victoria’s incidents grew by about 24% on a population that grew by only about 5%.

2. It is not one town

Chart 2. Each dot is a local government area (LGA), sized by population. Source: Crime Statistics Agency (2025), Tables 1-2; population implied by count divided by rate.

3. Where the climb is steepest

Chart 3. LGAs with at least 20,000 residents, ranked by percentage change in rate. Click a legend entry to isolate a line. Source: Crime Statistics Agency (2025), Table 2.

4. The twist: demand up, specialist help down

Specialist family violence homelessness services are one of the few places a victim-survivor can go when staying home is unsafe. The chart below compares, for each DFFH region, how much incidents grew with how much the number of people helped by specialist family violence homelessness services grew.

Chart 4. Each bubble is a DFFH region, sized by 2024-25 incidents. Source: Crime Statistics Agency (2025), Tables 1 and 17. Specialist clients are unique clients of specialist family violence homelessness services.

5. In 13 of 17 regions, specialist help per incident fell

Chart 5. Grey dot = 2020-21, coloured dot = 2024-25. A leftward move means fewer specialist clients for each incident recorded. This is a rough indicator of capacity relative to demand, not a coverage rate. Source: Crime Statistics Agency (2025), Tables 1 and 17.

What this means

Limits of this analysis

  • Incidents are police-recorded. A rise can reflect more reporting or better police response as well as more violence, so these data do not measure true prevalence.
  • Specialist clients are a single indicator of support. Other services (courts, health, counselling) are not shown, and cells with 3 or fewer people are suppressed by the CSA, so I counted them as 2.
  • Regions are Department of Families, Fairness and Housing (DFFH) regions; “Metro” and “Regional” groupings are my own classification.
  • Population is implied from the CSA’s counts and rates, and 2024-25 rates use Victoria in Future projections rather than final ABS estimates.
  • The analysis shows association over time, not cause.

Data, acknowledgements and references

Data. Crime Statistics Agency (2025) Data by Local Government Area, DFFH Region and Police Region, Data Tables, 1 July 2020 to 30 June 2025 [data set], Crime Statistics Agency Victoria, accessed 2026, https://www.crimestatistics.vic.gov.au/. Licensed under Creative Commons Attribution 4.0 International. Underlying sources: Victoria Police; Victorian Government (homelessness services data); ABS population estimates.

Generative AI acknowledgement. Generative AI [Claude] was used to assist with planning the story, writing R code, and drafting text. I reviewed, tested, and edited the generated output and take full responsibility for the final content.

Software. R Core Team (2025) R: A language and environment for statistical computing; Sievert C (2020) Interactive web-based data visualization with R, plotly, and shiny, CRC Press; Wickham H et al. (2023) dplyr, tidyr; Wickham H & Bryan J (2023) readxl.