The story

I came to Melbourne to study at RMIT and work night shifts in aged care. On a bus to work one night, a person threatened to stab me and shouted racist abuse; last year, my friend was stabbed on Elizabeth Street. Are assaults and weapon offences increasing in Melbourne, where do public transport offences occur, and how does the City compare with other councils? (Full story: “Why this story” tab.)

Assaults in the City of Melbourne

3,868

Weapon Offences in the City of Melbourne

1,135

Offences on Trains

159

Chart 1: Assaults and Weapon Offences: What Has Changed Since 2017?

City of Melbourne, 2017-2026. Use the menu to switch counts / rates; click the legend to show or hide offences.

Chart 2: Where Do Public Transport Offences Occur?

All offence types by location. Each panel has its own scale (from zero): compare trends; hover for counts.

Chart 3: How Does the City of Melbourne Compare with Other Councils?

Each dot is a council (hover for details). Dashed line = Victorian average rate. Orange = City of Melbourne.

Chart 4: Are Assaults Increasing Relative to Pedestrian Activity?

CBD assaults per million counted pedestrians (a proxy for people exposed). Grey = COVID-19 lockdowns.

Chart 5: When Is Melbourne Busiest, and What Does This Mean for Safety?

Press play to move through years. Shaded = night. Shows foot traffic, not when assaults occur.

Melbourne After Dark: Feeling Unsafe on the Way to Work

I came to Melbourne to study for my Master of Data Science at RMIT and build a better future for myself. Alongside my studies, I work night shifts as a kitchen hand in aged care, helping prepare and provide meals for residents. Because of my work, I sometimes have to travel at night, when there are fewer people around and getting to work can feel less safe.

On my way to work one night, I experienced a frightening incident on a Melbourne bus. A person threatened to stab me and shouted racist abuse at me. I was terrified and felt that I could be seriously hurt simply while travelling to work. The experience left me extremely distressed, and I almost had a panic attack. At that moment, I was not thinking about my studies or work; I was worried about my safety and whether I would reach my destination safely.

The incident was even more upsetting because my friend was stabbed on Elizabeth Street last year. Knowing that someone close to me had already experienced such violence made me even more concerned about my own safety. I began questioning whether these incidents were isolated events or part of a wider problem affecting people who live, study and work in Melbourne.

Working night shifts in aged care is a responsibility I take seriously. I want to do my job, support the residents and continue my education without being afraid of what might happen while travelling. Nobody should have to experience threats, racist abuse or fear of violence simply to get to work.

These experiences motivated me to investigate crime data in Melbourne. I wanted to understand whether assaults and weapon offences have increased since 2017, where public transport offences occur most frequently, and how the City of Melbourne compares with nearby suburbs and other Victorian councils.

By examining crime trends, locations and pedestrian activity, I aim to look beyond individual experiences and understand what the data tells us about public safety. Although statistics cannot capture the full emotional impact of an incident, they can help identify patterns and support informed discussions about how to make Melbourne safer for workers, students and commuters, particularly those who travel at night.

Why Night-Time Safety Matters to Me

I chose this topic because my personal experience of being threatened and subjected to racist abuse while travelling to work at night made me concerned about safety in Melbourne. My friend’s experience of being stabbed on Elizabeth Street last year made this concern even more personal.

As an international student at RMIT who works night shifts as a kitchen hand in aged care, I depend on public transport and need to travel to fulfil my work responsibilities. I believe everyone should be able to travel to work or return home without fearing violence or discrimination.

Rather than relying only on personal experiences or news reports, I wanted to investigate recorded crime data to understand how assaults and weapon offences have changed over time, where public transport offences occur, and how the City of Melbourne compares with other Victorian councils.

My aim is to use data to explore a public safety issue that affects workers, students, commuters and visitors. I also recognise that recorded crime figures cannot capture every incident or explain every person’s experience. By combining personal motivation with data analysis, I hope to contribute to a better understanding of safety in Melbourne, especially for people travelling at night.

Final takeaway

My experience of being threatened while travelling to work made me realise how important public safety is for people who work and study in Melbourne. By examining assaults, weapon offences, public transport locations and pedestrian activity, this project explores whether the data supports wider concerns about safety. The findings can help identify patterns and areas for further investigation, but they must be interpreted carefully. Everyone deserves to travel to work and return home without fear of violence or racist abuse.

About the data

Author: Ajith Kadavil Anilkumar (s4235429), Master of Data Science, RMIT University

This project investigates recorded crime trends and public safety in the City of Melbourne and other Victorian local government areas (LGAs), 2017-2026.

Research questions

1. Have recorded assaults and weapon offences increased since 2017?
2. Where are public transport-related offences most frequently recorded?
3. How does the City of Melbourne compare with other Victorian councils?
4. How does pedestrian activity vary by time of day, and what can this tell us about public safety?

Data sources

1. Crime Statistics Agency Victoria: recorded offences by LGA, suburb, offence type, location type (including street, train, tram and bus locations) and year, for the 12 months ending June 2017-2026.
2. City of Melbourne pedestrian counting system: hourly counts from pedestrian sensors in the city, used to measure activity across years and hours of the day.

Data preprocessing

• Crime data (CSA file, sheets Table 01-04): LGA names trimmed of extra spaces; offences grouped by subdivision code (A20 assault, A50 robbery, D10 weapons, C10 drug dealing, C30 drug use). Resident populations were back-calculated as offence count / rate per 100,000 x 100,000. CBD figures use the suburb “Melbourne” (postcode 3000).
• Pedestrian data (load_pedestrians.R): historic (2009-2022) and recent files combined, duplicate sensor-hours removed, dates converted to financial years (July-June) to match the crime data. Only sensors operating in every year were kept, so totals are not inflated by newly installed sensors.

How to read it: hover over any chart for exact values; click legend items to hide or show a series; use the buttons, search box and play button where shown; double-click to reset zoom.

Accessibility: colour-blind-safe Okabe-Ito palette, used consistently (orange = assault, blue = weapons); values are always available in hover text so colour is never the only cue; chart titles state the question in words.

Limitations

• Recorded offences are not the same as the number of people victimised. The data may be affected by reporting behaviour, population size, visitor numbers and changes in police recording practices.
• Raw offence totals do not necessarily indicate an individual’s risk of becoming a victim. Population-adjusted rates are useful for comparing councils, but the City of Melbourne has about 196,000 residents and far more daily visitors, so per-resident rates overstate individual risk. Public transport risk comparisons would ideally use passenger numbers.
• Location data covers all offence types combined, and the crime data has no time-of-day information.
• Pedestrian sensor counts approximate activity and do not represent every person in the city. A relationship between pedestrian numbers and assaults does not, by itself, establish that one causes the other.
• Each year is the 12 months ending 30 June, so 2026 is a full year (July 2025-June 2026), comparable with earlier years. 2020-2021 were affected by COVID-19 lockdowns.

Conclusion

This project uses recorded crime data to investigate whether violence-related offences have changed in Melbourne, where public transport offences occur and how the City of Melbourne compares with other councils. The analysis aims to provide evidence for a discussion about public safety while recognising the limitations of crime statistics and the importance of individual experiences.

References

City of Melbourne. (2026). Pedestrian counting system (counts per hour) [Data set]. City of Melbourne Open Data. Retrieved October 26, 2026, from https://data.melbourne.vic.gov.au/explore/dataset/pedestrian-counting-system-monthly-counts-per-hour/

Crime Statistics Agency. (2026). Data tables: LGA recorded offences, year ending June 2026 [Data set]. https://files.crimestatistics.vic.gov.au/2026-09/Data_Tables_LGA_Recorded_Offences_Year_Ending_June_2026.xlsx

Kumar, G. P. (2022, July 6). HR attrition dashboard [Data visualisation; design inspiration]. Tableau Public. https://public.tableau.com/app/profile/pradeepkumar.g/viz/HRAttritionDashboardRWFD_16570446563570/viz

R packages: flexdashboard, plotly, ggplot2, dplyr, stringr, readxl, crosstalk, data.table. (Run citation("package") in R for each to get its APA reference.)