I moved to Melbourne in February 2026 as an international student, and most things here still feel new: the trams, the myki card, half the suburb names I can’t pronounce yet. Part of settling into a new city is figuring out what can go wrong, so for this assignment I opened up the Victorian crime statistics expecting the usual story about stolen cars. What I found instead was smaller, and a bit odd.
In the year to June 2026, police recorded 32,134 stolen number plates in Victoria. That works out to roughly 88 a day, and it’s 124% more than the 14,363 recorded four years earlier, back when I was still living in India. A plate on its own is worth nothing. People steal them to put on another car so it can’t be traced, whether that’s for a petrol drive-off, dodging tolls, or something worse like a ram raid (RACV, n.d.). Here’s what the data shows, across five charts.
Quick tip: hover over any chart for exact numbers, click a legend item to hide or show it, and drag across a chart to zoom in (double-click to zoom back out).
Data: Crime Statistics Agency (2026a), Property Items Table 06 (number-plate thefts) and Recorded Offences Table 01 (subgroups B41, B42). “Other theft from cars” = all steal-from-motor-vehicle offences minus those recorded as number-plate theft. Years end 30 June. Use the buttons above the chart to switch between raw counts and growth since 2022.
Follow the orange line. It dips during the lockdowns, which makes sense since fewer people were out and about, and then from 2023 it doesn’t stop climbing. Whole-car theft rises at nearly the same pace (more on that in chart 3), but plates grew fastest of all three and now make up 42% of everything stolen from parked cars. Click the index button and you can see how far they’ve pulled ahead of other theft from cars.
Data: Crime Statistics Agency (2026a) Property Items Table 06; populations derived from Crime Statistics Agency (2026b) LGA Table 01 (offence count ÷ rate per 100,000). I’ve shown the seven local government areas (LGAs) in each police region with the highest 2026 rate, skipping any LGA with fewer than 20,000 residents. Grey dots are 2022 and orange dots are 2026. All four panels share the same x-axis, so you can compare regions directly.
I assumed this would mostly be an outer-suburbs problem. It isn’t. Maribyrnong and Port Phillip, both close to the city, sit right alongside Greater Dandenong at the top. Casey had the biggest jump in raw numbers, going from 772 stolen plates to 2,754. It’s not just Melbourne either: Ballarat and Geelong went up too.
Data: Crime Statistics Agency (2026a) Property Items Table 06; Crime Statistics Agency (2026b) LGA Tables 01 and 02 (subgroup B41 Motor vehicle theft). Only LGAs with more than 20,000 residents are included. The left panel is 2022 and the right panel is 2026. Bubble size shows population. Double-click a region in the legend to show just that region.
Compare the two panels: between 2022 and 2026 the bubbles all move up and to the right together. One crime didn’t replace the other. The areas losing more cars were losing more plates as well. That lines up with what local police have said, that stolen plates get used to disguise a car and avoid being picked up (Star Mail, 2025). The data alone can’t prove a link, but the pattern is pretty consistent.
Data: Crime Statistics Agency (2026a) Property Items Table 06. Shows the 10 suburbs with the most number-plate thefts in 2026. Dandenong is in orange; hover over a grey line to see which suburb it is.
32,000 plates across the state is hard to picture, so this one zooms in on the suburbs hit hardest. When I first arrived I couldn’t tell one suburb from another, so this was a useful way to see where it’s worst. Dandenong topped the list in 2026, but every one of these lines bends upward after 2022.
Data: Crime Statistics Agency (2026a) Recorded Offences Table 04 (investigation status when the data were extracted) and Table 05 (offence codes). The CSA doesn’t publish outcomes for plate thefts separately, so I’ve used the whole steal-from-motor-vehicle group instead. The black tick marks the unsolved share in 2017. Click a legend item to hide that outcome.
This was the part that surprised me most. 94% of thefts from cars in 2026 were never solved, a higher share than burglary, bike theft or stolen cars. In that same year, police across the entire state recorded just 963 offences for using, forging or altering a number plate or rego label, about one for every 33 plates stolen. In other words, very few people are ever charged for using a stolen plate.
It’s a small crime, and easy to overlook next to the bigger headlines. But it’s growing quickly, it’s showing up all over Melbourne and regional Victoria, and hardly anyone gets caught. The upside is that protecting yourself is cheap. One-way anti-theft screws cost around $10, and Neighbourhood Watch runs “Safe Plate” events, backed by RACV and Victoria Police, where they’ll fit them for a gold coin donation (RACV, n.d.). It’s also a good idea to blur your plates in any photos of your car you put online. And if a plate does go missing, report it to 131 444 straight away so you don’t get stuck with someone else’s fines and tolls.
All data comes from the Crime Statistics Agency (CSA) Victoria release for the year ending June 2026 (published September 2026), and the code pulls it straight from the CSA website. Number-plate thefts are steal-from-motor-vehicle offences where a number plate was recorded as stolen (Property Items Table 06). Statewide totals use the CSA’s “Victoria” rows. LGA rows exclude a small number of offences with unknown location. Rates per 100,000 use populations derived from the CSA’s own LGA rates, which are based on ABS and Victorian in Future estimates. It’s worth remembering that this is recorded crime, so it only includes what was reported to police and recorded by them. The CSA also notes some data-quality issues with the property-items variable, so the true numbers are likely higher. Investigation status is the status when Victoria Police extracted the data (the 18th day after the reference period) and can change later.
Anthropic. (2026). Claude (Opus 5.5) [Large language model]. https://claude.ai
Baglin, J. (n.d.). Data visualisation: From theory to practice. Retrieved October 27, 2026, from https://data-visualisation.stem.melbourne/
Crime Statistics Agency. (2026a). Recorded offences visualisation and property items visualisation data tables, year ending June 2026 [Data set]. Victorian Government. https://www.crimestatistics.vic.gov.au/crime-statistics/latest-victorian-crime-data/download-data
Crime Statistics Agency. (2026b). LGA recorded offences data tables, year ending June 2026 [Data set]. Victorian Government. https://www.crimestatistics.vic.gov.au/crime-statistics/latest-victorian-crime-data/download-data
Okabe, M., & Ito, K. (2008). Color universal design (CUD): How to make figures and presentations that are friendly to colorblind people. https://jfly.uni-koeln.de/color/
R Core Team. (2025). R: A language and environment for statistical computing (Version 4.5.2) [Computer software]. R Foundation for Statistical Computing. https://www.R-project.org/
RACV. (n.d.). How to help protect your car from number plate theft. RoyalAuto. Retrieved October 27, 2026, from https://www.racv.com.au/royalauto/transport/car-safety-security/rise-in-number-plate-theft.html
Sievert, C. (2020). Interactive web-based data visualization with R, plotly, and shiny. Chapman and Hall/CRC. https://plotly-r.com
Star Mail. (2025, July 13). Secure your plates: Number plate thefts surge. Upper Yarra Star Mail. https://upperyarra.mailcommunity.com.au/news/2025-07-13/secure-your-plates-number-plate-thefts-surge/
Wickham, H. (2016). ggplot2: Elegant graphics for data analysis (2nd ed.). Springer. https://ggplot2.tidyverse.org
Data are © State of Victoria (Crime Statistics Agency), licensed under a Creative Commons Attribution 4.0 International licence.
GenAI Acknowledgement
I acknowledge the use of Claude (Anthropic, 2026; https://claude.ai) to help fix R coding errors, check spelling and grammar, and draft this acknowledgement. Some prompts are mentioned below:
“I’m knitting an R Markdown report in RStudio and one of my plots won’t render. When I run this chunk on its own, I get an error:” “Write me an AI acknowledgement for this”
I used the output only as guidance. I checked it, edited it and applied it to my own work. I did not use AI to come up with my story idea, write content or create any visualisations.
The data is © State of Victoria (Crime Statistics Agency) and licensed under CC BY 4.0. The charts were built in R using readxl, dplyr, tidyr, stringr, ggplot2 and plotly. The colours are adapted from the Okabe and Ito (2008) colour-blind-safe palette, and the design and interactivity choices follow Baglin (n.d.) and Sievert (2020).