Skip to content

safety

Sydney crime statistics by suburb, read properly

Sydney crime statistics by suburb

BOCSAR's suburb data for 40 Sydney suburbs, with the police-detected offences split out so the numbers mean what they seem to mean.

By Sydney Security QuotesReviewed

The short answer

Across the 40 Sydney suburbs we track, NSW Police recorded 103,615 criminal incidents in the year to June 2026. More than half of them (53%) were offences police find themselves, such as fare offences at stations, bail breaches near courts and drug possession. Those numbers follow police activity, not risk to a home or business.

Strip them out and 48,258 offences against people and property remain, up 3.1% on the year before and 4.6% on three years earlier. Inside that, the change that matters most for businesses is theft from shops: 9,121 recorded incidents, up 15.5% in a year and 20.3% in three. Trespass is up 51.8% over three years. Break-ins to commercial premises are down 7.8% on the year before.

Where these numbers come from

The NSW Bureau of Crime Statistics and Research (BOCSAR) publishes every criminal incident recorded by NSW Police, by suburb and month, back to 1995. BOCSAR updates the data every quarter, about three months after each quarter ends. This page uses the suburb file released in September 2026, which runs to June 2026. The next release is due in December 2026, and we refresh this page from each one.

Every figure here covers July 2025 to June 2026 unless it says otherwise. We downloaded the file on 15 November 2026.

A few rules of the data:

  • Recorded is not the same as committed. BOCSAR counts incidents reported to or found by police. Some offences, such as shop theft and assault, are under-reported, and reporting can change when a retailer or a venue changes its policy.
  • The counting unit is the incident. One incident can involve several offences of the same type. Murder and manslaughter are the exception: BOCSAR counts victims.
  • Counts follow the place of the incident, not where the offender or victim lives.

Why the total misleads

Police-detected offences inflate busy centres. Liverpool recorded 14,574 incidents, more than any of the 40 suburbs and far more than the Sydney CBD. But 8,205 of them were transport regulatory offences, which BOCSAR describes as offences on the rail network such as travelling without a valid ticket, and 80% of Liverpool's total was police-detected. Against people and property, Liverpool recorded 2,954 offences, a little over half the CBD's 5,131. Parramatta shows the same pattern: of its 9,371 incidents, 1,419 were transport offences and 1,188 were breaches of bail conditions, and 60% of the total was police-detected.

Raw counts punish busy places. The Sydney CBD and Parramatta have daytime populations many times their resident populations. A suburb that hosts huge foot traffic records high counts without any one person or premises facing unusual risk. That's also why we don't divide by population: a CBD's residents are a small share of the people in it.

The five-year window still starts in a lockdown. July 2021 to June 2022 includes the 2021 Sydney lockdown. That year BOCSAR's "other offences" jumped: 3,017 in the Sydney CBD, against about 300 in each year since. The jump lines up with public health order breaches, which BOCSAR separately counted at 36,597 across NSW between 26 June and 31 August 2021. Shop theft was unusually low at the same time. So five-year trends that start there overstate falls in the total and rises in theft. Until the window moves past mid 2022, compare against three years ago instead. The table does both.

The 40 suburbs: offences against people and property

SuburbPeople and property offencesChange on the year beforeChange over 3 years5-year trendAll recorded incidents
Alexandria677-16.1%+24.7%Up 7.7% a year846
Artarmon398+14.4%+22.1%Up 9.4% a year582
Ashfield884-5.8%+5.1%Up 3.3% a year1,429
Auburn1,580-1.8%-0.7%Up 7% a year2,758
Bankstown1,929-8.2%-5.7%Up 4.8% a year5,093
Blacktown3,421+0.1%-12.1%Stable8,114
Bondi Junction1,109-5.5%-24%Stable1,839
Botany299+0.3%-9.4%Stable384
Burwood1,341+3.2%+6.6%Up 8.3% a year2,499
Castle Hill969-7.4%-12.8%Stable1,438
Chatswood1,418+5.3%+15.8%Up 10.1% a year3,330
Crows Nest305+17.3%+54%Up 11.5% a year459
Dee Why582+2.3%-21.2%Stable895
Double Bay236+11.3%+7.8%Stable417
Hornsby876+4.8%+1.7%Stable2,660
Hurstville1,612+13.2%+26.8%Up 14.9% a year3,923
Kogarah583-5.7%+1.7%Stable1,442
Liverpool2,954-5.7%-2.1%Up 4.2% a year14,574
Macquarie Park933+11.6%+47.2%Up 16.8% a year1,171
Manly788-9.6%-9.7%Stable1,308
Marrickville1,609+19.7%+24.5%Up 10.3% a year2,568
Mascot994-2.3%+2.3%Up 6.7% a year1,410
Newtown1,169+19.5%+15.1%Up 4% a year2,383
North Sydney354+11.7%+18.4%Stable584
Parramatta3,765+5.8%+11.9%Up 9% a year9,371
Penrith3,000+3%+13.1%Up 5.6% a year6,273
Pyrmont714+8.3%-3.9%Stable1,374
Randwick1,101+21.5%+4.1%Up 7.4% a year1,620
Rhodes515+5.3%+27.8%Up 9.7% a year651
Ryde869+3.1%+7%Up 7.4% a year1,341
Silverwater309+3%+8.4%Up 10.1% a year354
Smithfield533-4.5%-0.4%Stable685
St Leonards322+22%+45%Up 15% a year677
Strathfield1,122+7.6%+6.7%Up 6.4% a year1,801
Surry Hills1,808+3.3%+0.3%Up 7.2% a year4,360
Sutherland439-0.7%-4.1%Stable1,797
Sydney CBD5,131+7.1%+6.3%Up 9.1% a year8,638
Sydney Olympic Park418-0.2%+14.5%Up 8.1% a year1,187
Wetherill Park546+5.6%+15.2%Up 13.9% a year639
Zetland646+18.1%+27.9%Up 9.9% a year741

Source: NSW Bureau of Crime Statistics and Research (BOCSAR), recorded criminal incidents, July 2025 to June 2026.

How to read it:

  • People and property offences are offences with a victim: assault, sexual offences, robbery, theft, break and enter, fraud, malicious damage, arson, intimidation and the like. They leave out police-detected offences (drugs, weapons, disorderly conduct, liquor, transport and justice offences, receiving stolen goods and assaults on police).
  • Change on the year before compares July 2025 to June 2026 with July 2024 to June 2025.
  • Change over 3 years compares it with July 2022 to June 2023, the first full year after the 2021 lockdown.
  • 5-year trend uses BOCSAR's own test: Kendall's tau-b on the 60 monthly counts from July 2021, with a result reported only when it is statistically significant (p < 0.05). "Stable" means no significant rise or fall. Read it together with the lockdown note above.
  • All recorded incidents is BOCSAR's full count, including police-detected offences.

The offences that matter most to premises

SuburbSteal from retail storeBreak and enter (non-dwelling)Malicious damageNon-domestic assaultTrespass
Alexandria9617603621
Artarmon3430473434
Ashfield12581029637
Auburn983515417755
Bankstown1872418520151
Blacktown53941374435153
Bondi Junction46319808242
Botany3242196
Burwood379399711035
Castle Hill24813736723
Chatswood5203310614339
Crows Nest83630308
Dee Why7110656920
Double Bay29624392
Hornsby212161128333
Hurstville5612015211686
Kogarah405456312
Liverpool37830345343131
Macquarie Park33616496921
Manly3777913263
Marrickville2993020310073
Mascot9516848110
Newtown1362015317633
North Sydney7115313421
Parramatta1,01361292399308
Penrith75046356328206
Pyrmont318797038
Randwick196131209355
Rhodes702237266
Ryde10331106319
Silverwater17310144
Smithfield1215504113
St Leonards5012333017
Strathfield207128211256
Surry Hills834318925198
Sutherland3511814024
Sydney CBD1,36086432895261
Sydney Olympic Park148387634
Wetherill Park768605612
Zetland647421714

Source: NSW Bureau of Crime Statistics and Research (BOCSAR), recorded criminal incidents, July 2025 to June 2026.

These five offence types are the ones a security provider can actually change:

  • Steal from retail store is shop theft. It concentrates where the shopping centres are: the Sydney CBD (1,360), Parramatta (1,013), Penrith (750), Hurstville (561), Blacktown (539) and Chatswood (520).
  • Break and enter (non-dwelling) is a break-in to a shop, office, warehouse, school or site.
  • Malicious damage covers vandalism and graffiti.
  • Non-domestic assault is assault outside a family or household setting, which includes assaults in venues, shops and streets.
  • Trespass is being on premises without a lawful reason, often the first step before a break-in.

What changed across Sydney this year

Totals for the 40 suburbs, July 2025 to June 2026:

OffenceIncidentsChange on the year beforeChange over 3 years
Offences against people and property48,258+3.1%+4.6%
Steal from retail store9,121+15.5%+20.3%
Non-domestic assault5,246+2.9%+9.6%
Malicious damage4,703+0.9%-0.2%
Steal from motor vehicle2,562-6.7%-8.2%
Trespass2,174+15.5%+51.8%
Break and enter (dwelling)1,496-8.3%-3.3%
Motor vehicle theft1,390-1.3%+18.3%
Break and enter (non-dwelling)816-7.8%+1.9%

Source: NSW Bureau of Crime Statistics and Research (BOCSAR), recorded criminal incidents, July 2025 to June 2026. Totals add up BOCSAR's counts for the 40 suburbs.

Over the last 24 months, BOCSAR's test finds a statistically significant rise in offences against people and property in five suburbs: Marrickville, Newtown, Randwick, St Leonards and the Sydney CBD. It finds a significant fall in two: Alexandria and Bankstown. Everywhere else the year-on-year movement is within normal variation.

Crime maps for Sydney

The official crime map for NSW is BOCSAR's Crime Mapping Tool. It's free, covers the whole state, and is built on the same recorded crime data as the tables on this page. NSW Police points people there too: its crime statistics page sends anyone after figures for their local area to BOCSAR.

The tool has two kinds of map, plus table, trend and chart views:

  • Rate maps, the default view, shade each area by recorded crime per 100,000 residents. You can switch between local government areas, postcodes, suburbs and larger SA4 regions.
  • Hotspot maps show where incidents cluster, down to street level, for 11 offences: domestic assault, non-domestic assault, alcohol-related assault, robbery, break and enter on homes, break and enter on other buildings, motor vehicle theft, theft from vehicles, theft from a person, theft from homes, and malicious damage. Shop theft and trespass, the two biggest risers above, have no hotspot maps.

Counts go back to the mid 1990s and suburb rates to 2012, and the tool is updated every three months, with BOCSAR's quarterly release. It currently runs to June 2026.

Looking up one suburb:

  1. Open the tool and the menu at the top left.
  2. Under Boundary, choose Suburb.
  3. Type the suburb into the Location box and pick it from the Suburb results, not the council area of the same name.
  4. Choose the offence. The default period is the latest two years of data.
  5. Open the Table view. The suburb is pinned under NSW, with its count, its rate, how many times the NSW rate that is, and its two-year trend.
  6. For the 11 hotspot offences, open the Hotspots Map tab to see where in the suburb incidents cluster.

Reading the maps without being misled:

  • A hotspot is a cluster, not a risk score. BOCSAR lays a 50 metre grid over NSW, scores every cell for the density of nearby incidents, and shows the densest cells as red, orange and yellow hotspots. Together they usually cover less than 1% of the state's land area. Hotspots aren't adjusted for how many people live in or pass through an area, so shopping strips, stations and nightlife streets stand out because that's where the people are. A few incidents close together can make a hotspot, and a hotspot inside a suburb doesn't mean the suburb has a high count.
  • Rates per resident punish busy centres. BOCSAR warns that the Sydney local government area has high recorded crime rates because its resident population is small. The same goes for any centre full of workers and shoppers, which is why the tables on this page use counts and offence types instead.
  • Most suburbs have no rate at all. BOCSAR doesn't calculate a rate or a trend for suburbs with fewer than 2,000 residents, and it says that leaves about 80% of NSW suburbs without one. On the map they're hatched or left clear.
  • It's recorded crime, mapped from police records. Crimes nobody reports aren't on it, and offences police detect themselves follow where police are working. Each incident is mapped from the location in the police record, so the map is only as precise as that record, and BOCSAR doesn't publish maps of single incidents.

For a decision about a site, use the map to see where in a suburb break-ins, theft from vehicles or damage cluster. Then compare suburbs with the tables above, which split out police-detected offences and include shop theft and trespass.

What the pattern means for a security decision

Match the offence to the cover, not the suburb to a fear:

Shop theft during trading hours points to a uniformed guard at the entry through peak trade, or a plain-clothes loss prevention officer where losses are concentrated. See retail security.

Trespass and after-hours break-ins point to mobile patrols with alarm response, so an alarm gets a licensed officer on site instead of you.

Malicious damage is usually a patrol, lighting and camera question rather than a guard question.

Assaults around venues and late trade sit with venue security and crowd control on the nights that need it.

One exposed point, such as a gate, dock or fit-out floor, points to a static guard for the risky hours.

Suburb data is a useful starting point and a poor substitute for someone walking your actual site. Run the estimator for your site and hours, and a licensed provider confirms the real price.

Method

Source. NSW Bureau of Crime Statistics and Research (BOCSAR), Recorded Criminal Incidents by month, by Suburb, July 2025 to June 2026, accessed 15 November 2026. © State of New South Wales (Department of Communities and Justice), licensed under CC BY 4.0. We selected 40 suburbs, grouped the offences and added up the monthly counts.

  • Privacy adjustment. BOCSAR says it randomly adjusts the counts it publishes, totals included, to protect privacy, and that adding adjusted counts together can introduce small errors. The yearly and 40-suburb totals here add up BOCSAR's monthly suburb counts, so read small numbers and totals as close, not exact.
  • Suburbs. The 40 suburbs on this site, matched to BOCSAR's gazetted suburb names. "Sydney CBD" is BOCSAR's suburb "Sydney". Haymarket, The Rocks, Barangaroo, Millers Point and Dawes Point are separate suburbs and not included. Silverwater is "Silverwater (Parramatta)".
  • Groups. People and property offences are the BOCSAR categories with a victim, less receiving stolen goods and assaults on police, which police mostly detect. The full list of categories is in the data file.
  • Trends. Kendall's tau-b on monthly counts, two-tailed, p < 0.05, as BOCSAR describes on its "Using crime statistics" page. No test where any 12-month total is under 20. Where the trend is significant, the average annual change is the compound rate between the first and last 12-month totals.
  • What we don't do. We don't estimate, adjust, model or forecast, and we don't calculate rates per person. Where BOCSAR has no figure, we show nothing.
  • Refresh. Rebuilt from each quarterly BOCSAR release.

Questions

Is crime going up in Sydney?

Across these 40 suburbs, offences against people and property rose 3.1% in the year to June 2026. Shop theft and trespass rose much faster. Break-ins to homes and businesses and theft from cars fell.

Which Sydney suburb has the most crime?

Of these 40, Liverpool records the most incidents in total, but 80% are police-detected, mostly rail network offences such as travelling without a valid ticket. Against people and property, the Sydney CBD records the most (5,131), then Parramatta (3,765) and Blacktown (3,421). All three are major centres with large daytime populations.

Why don't you show crime rates per person?

Rates divide by residents, and in commercial centres most people present are workers, shoppers and visitors. A per-resident rate would make every CBD look dangerous and every dormitory suburb look safe.

Is there a crime map for Sydney?

Yes. BOCSAR's Crime Mapping Tool maps recorded crime across NSW by council area, postcode and suburb, with rates per 100,000 residents where an area has enough residents, and hotspot maps for 11 offences. It's free at crimetool.bocsar.nsw.gov.au and updated every three months.

What do the hotspots on the map mean?

That incidents of that offence are densely clustered there. BOCSAR scores 50 metre grid cells across NSW for the density of incidents, and the hotspots usually cover less than 1% of the state's land area. Hotspots don't allow for how many people are in an area, so they show where incidents concentrate, not how likely you are to be a victim.

How often is this updated?

Every quarter, after BOCSAR's release. The period and the access date are always shown with the figures.

Sources

One minute to your quotation

Answer a few questions, get your number, and see what to confirm next.

Get my quotation