I work the register at a supermarket in Parkville. Most weeks I watch someone walk out without paying, usually for alcohol or food. Sometimes I get the item back. Sometimes the person turns aggressive and I let it go.

Victoria’s crime data lets us check how that experience compares with the official record.

Supermarkets are where shoplifting grew fastest

Recorded thefts from Victorian supermarkets more than doubled in nine years. They fell during the COVID lockdowns, then climbed every year since.

Try it: switch between the number of thefts and the change since 2017. Click a store type in the legend to hide or show it.

Source: Crime Statistics Agency (2026a), Recorded offences, Table 02. Offence: B43 Steal from a retail store, by location type. Years end 30 June.

The city has always been the hotspot. Now the rest of Victoria is catching up

The City of Melbourne, where I live and work, records more shop theft per resident than any other council, as it did in 2017. What has changed is everywhere else: Yarra’s rate more than doubled and Port Phillip’s quadrupled.

Try it: hover over a council to see the number of offences behind each rate.

Source: Crime Statistics Agency (2026b), Recorded offences by LGA, Table 02. Rates are per resident, so councils with many visitors and workers (like the CBD) appear higher.

But where I work, the thefts don’t show up

Across the suburbs of the City of Melbourne, the CBD lights up every year. Parkville, where I work, is almost blank: 12 recorded shop thefts in ten years.

Try it: hover over any square for the exact count. Darker means fewer recorded thefts.

Source: Crime Statistics Agency (2026b), Recorded offences by LGA, Table 03 (suburb level). Offence: B43 Steal from a retail store. Colour uses a log scale.

That near-empty row doesn’t match my shifts. When someone steals at my store, staff report it to the manager. Whether it ever reaches the police is out of our hands, and only what reaches the police becomes a number in this data. The gap between what happens at the counter and what gets recorded is the part of this story the statistics can’t show.

The people being caught are not who you’d expect

The picture of the teenage shoplifter doesn’t match the data. Since 2017, the largest growth in people accused of property crime is among adults in their 30s and early 40s.

Try it: press Play or drag the slider to watch the shape change from 2017 to 2026. Hover for exact counts.

Bars show alleged offender incidents for property and deception offences: males to the left, females to the right. Source: Crime Statistics Agency (2026c), Alleged offender incidents, Table 04. Covers all property and deception offences (shop theft is not published separately by age). Counts are incidents, not unique people.

And the job itself is getting more dangerous

Theft is only part of it. Assaults, weapons, drug and robbery offences recorded at supermarkets have all risen. These are the moments when staff have to decide whether to step in or let it go.

Try it: hover over a point for the exact count. Each panel has its own scale so the trend is visible for rarer offences.

Source: Crime Statistics Agency (2026a), Recorded offences, Table 02. Location: 335 Supermarket. Offence subdivisions A20 Assault, A50 Robbery, C30 Drug use and possession, D10 Weapons.

What the numbers miss

The official record says supermarket shoplifting has never been higher. From behind the counter, even that sounds like an undercount. Much of what I see stops at the manager’s office, and never becomes a number: the bag that walks out the door, the item I get back, the one I let go because the person turned aggressive. The data can tell us how often theft is reported and who gets caught. It can’t show how often staff are left to make that call on their own, in a few seconds, at the checkout.

About the data

  • Recorded offences are crimes reported to or detected by Victoria Police and recorded on its database. Shop thefts that are never reported to police are not included, so these figures are a minimum.
  • Shop theft means the Crime Statistics Agency offence subgroup B43 Steal from a retail store. Charts 1 and 5 use the CSA location classification to identify offences at supermarkets (335 Supermarket).
  • All figures are for years ending 30 June. Rates in Chart 2 are per 100,000 residents and do not account for the workers, students and visitors in each area.
  • Chart 4 counts alleged offender incidents for all property and deception offences, because the CSA does not publish shop theft separately by age and sex. One person can be counted more than once.
  • Data were downloaded from DataVic as Excel workbooks and processed in R using readxl, dplyr, tidyr and stringr. Charts were made with plotly and ggplot2. All code is included in the submission.

AI Acknowledgement

Generative AI (Chatgpt) was used to assist with refining written explanations, reviewing spelling and grammar, and providing guidance on using R and data visualisation tools. All suggestions were reviewed and verified for accuracy and relevance before being incorporated into the assignment. I take full responsibility for the final content and submitted work.

References

Crime Statistics Agency. (2026a). Recorded offences – Year ending June 2026 [Data set]. DataVic. https://discover.data.vic.gov.au/dataset/data-tables-recorded-offences/resource/35098a5b-d906-446f-a7fb-eaf07270ce45

Crime Statistics Agency. (2026b). Recorded offences by LGA – Year ending June 2026 [Data set]. DataVic. https://discover.data.vic.gov.au/dataset/data-tables-recorded-offences/resource/06bb24c8-8f62-4144-a746-99ef5492e4f9

Crime Statistics Agency. (2026c). Alleged offender incidents – Year ending June 2026 [Data set]. DataVic. https://discover.data.vic.gov.au/dataset/data-tables-alleged-offender-incidents/resource/202292f5-c813-4878-8e43-153825e87670

OpenAI. (2026). ChatGPT [Large language model]. https://chatgpt.com/