Cafés and restaurants are an important part of Melbourne’s urban environment. Their seating capacity offers one way to examine how the city’s dining landscape has changed over time. However, a city-wide total can conceal differences between neighbourhoods, and changes in recorded seating capacity do not necessarily correspond directly to changes in the number of businesses.
This report examines recorded dining capacity in the City of Melbourne between 2002 and 2024 using the City of Melbourne’s Café, restaurant, bistro seats dataset. The analysis focuses on records classified under ANZSIC code 4511, Cafes and Restaurants, and distinguishes between indoor and outdoor seating.
The analysis addresses three questions:
The three interactive visualisations present the long-term trend, geographic differences, and relationship between changes in property IDs and seats. The findings are descriptive and do not establish the causes of the observed changes.
The dataset is downloaded from the City of Melbourne Open Data Portal. The analysis is restricted to ANZSIC 4511 records with a valid census year and non-negative recorded seat count.
Figure 1. Recorded indoor and outdoor dining seats in the
City of Melbourne, 2002–2024.
Source: City of Melbourne (n.d.).
Both indoor and outdoor seating increased substantially over the earlier part of the period, reached higher levels in the mid-to-late 2010s, and then declined. The data show a partial recovery in 2024, but the total remained below the 2019 level.
Separating indoor and outdoor seating makes it possible to examine whether these components followed similar patterns. Between 2019 and 2024, indoor seating changed by -13,761 seats (-11.42%), while outdoor seating changed by -3,704 seats (-13.63%). The absolute decline was greater for indoor seating, while outdoor seating experienced the larger percentage decrease. The interactive chart allows readers to inspect individual years and compare the two series using the legend and hover information.
The vertical reference line at 2020 provides temporal context for examining the period around the COVID-19 pandemic. However, the chart does not establish that the pandemic caused the decline. Changes in recorded seating capacity may have several explanations, and further evidence would be needed to identify the factors responsible.
Figure 2. Change in recorded dining seats by City of
Melbourne small area, 2019–2024.
Source: City of Melbourne (n.d.).
The decline was not distributed evenly across the municipality. In the comparison shown, recorded seats fell in 10 of 12 areas. Docklands, Melbourne (CBD) and East Melbourne had the largest reductions. West Melbourne (Residential) and Kensington recorded small increases.
The three largest reductions totalled 14,613 seats. These differences show why a city-wide total should not be interpreted as a uniform experience across all areas.
The comparison uses areas with valid records in both 2019 and 2024. It describes the difference between two endpoints rather than every change that occurred in between. The observed differences should not automatically be attributed to business closures or changes in demand: the dataset alone cannot distinguish between changes in seating arrangements, establishments, business activity or recording.
Figure 3. Relationship between changes in distinct recorded
property IDs and recorded seats, 2019–2024.
Source: City of Melbourne (n.d.).
The scatter plot compares changes in distinct recorded property IDs with changes in seating capacity. Areas to the left of zero had fewer distinct recorded property IDs in 2024 than in 2019, while areas below zero had fewer recorded seats. Areas in the lower-left quadrant therefore experienced reductions in both measures.
The measures are not interchangeable. A property can accommodate different numbers of seats, and property IDs do not necessarily represent individual cafés or restaurants. The scatter plot shows whether the measures changed together; it does not establish that changes in property IDs caused changes in seating capacity.
The three visualisations provide complementary perspectives on changes in recorded dining capacity in the City of Melbourne. Recorded seating capacity declined from its late-2010s levels and remained below the 2019 total in 2024, despite a partial increase in the final year. The geographic comparison shows that the decline was uneven across small areas, while the scatter plot compares changes in seating capacity with changes in distinct recorded property IDs.
These findings describe recorded seating capacity rather than the complete performance of Melbourne’s hospitality sector. The data does not directly measure customer demand, revenue or business survival, and property IDs are not equivalent to individual businesses. The analysis also does not establish the causes of the observed changes.
Further research could combine seating records with independent data on business openings and closures, employment, trading conditions or other measures of hospitality activity. This would help investigate whether changes in recorded seating capacity reflect changes in the number of establishments, changes in seating arrangements or other factors.
The analysis uses the City of Melbourne’s Café, restaurant, bistro seats dataset, which contains records from 2002 to 2024. The source describes establishment trading names, addresses, industry classifications, seating types, seat counts and locations. The analysis is limited to the City of Melbourne municipality and should not be interpreted as representing all suburbs of Greater Melbourne.
The raw data was loaded into R and cleaned by converting census year and seat count fields to numeric values, standardising text fields, and retaining ANZSIC 4511 (Cafes and Restaurants) records with valid census years and non-negative seat counts. Annual seat totals were calculated by year and seating type. For the geographic comparison, recorded seats were summed by small area for 2019 and 2024, and the difference was calculated. For the property comparison, distinct property IDs were counted by small area and year and compared with changes in recorded seats.
The visualisations were created in R using tidyverse,
plotly and scales. The interactive charts
provide hover information; the third chart also provides buttons to
switch between the full range and a zoomed view.
The results should therefore be interpreted as a descriptive analysis of the recorded dataset rather than a complete explanation of changes in Melbourne’s hospitality sector.
Generative AI assistance was used to help debug R code and organise the report. I reviewed and edited the drafted text to make sure it accurately reflects my own reasoning and the actual results from my code before including it here. AI use is acknowledged in accordance with the course guidance.
City of Melbourne. (n.d.). Café, restaurant, bistro seats [Data set]. City of Melbourne Open Data Portal. https://data.melbourne.vic.gov.au/explore/dataset/cafes-and-restaurants-with-seating-capacity/
OpenAI. (2026). ChatGPT [Large language model]. https://chatgpt.com/
R Core Team. (2026). R: A language and environment for statistical computing. R Foundation for Statistical Computing. https://www.R-project.org/
Sievert, C. (2020). Interactive web-based data visualization with R, plotly, and shiny. Chapman and Hall/CRC. https://plotly-r.com/
Wickham, H., Averick, M., Bryan, J., Chang, W., McGowan, L. D., François, R., Grolemund, G., Hayes, A., Henry, L., Hester, J., Kuhn, M., Pedersen, T. L., Miller, E., Bache, S. M., Müller, K., Ooms, J., Robinson, D., Seidel, D. P., Spinu, V., … Yutani, H. (2019). Welcome to the tidyverse. Journal of Open Source Software, 4(43), Article 1686. https://doi.org/10.21105/joss.01686