Showcase · Data layers · City of Melbourne open data
The CBD changing use
Between 2002 and 2024, apartment floor space in Melbourne’s CBD grew 7.4 times, from 0.31 to 2.28 million m². Office floor space ended about where it began: 2.75 million m² in 2002, a peak of 3.19 million in 2008, 2.83 million in 2024. Block by block, where the census publishes both years, 9 CBD blocks lost over 5,000 m² of office and gained over 5,000 m² of apartments, hotels or student rooms: 143,008 m² of office out, 365,305 m² in.
Office flat, apartments up
The council’s Census of Land Use and Employment (CLUE) measures every block’s floor space by what it is used for. Across the 83 CBD blocks, office grew to 3.19 million m² by 2008, held there to 2019, then fell back to 2.83 million by 2024. Apartments took off after 2012 and passed 2 million m² in 2022. Hotels and serviced apartments went from 0.23 to 0.43 million m². Student housing in the CBD is almost all held back by the census (only 3,033 m² is published for 2024), so it can’t be followed here.
The map
Each block coloured by apartments or office as a share of its floor space, for the census year on the slider, or by the change in office from 2002 to 2024. Press Play to step through the years. Hover over a block, or tap it on a phone, for its numbers and its office and apartments over the years. A pale block with a dashed edge has its figure held back by the census.
Office out, homes and hotels in
Three blocks on Collins Street lost the most office: Collins St, William St to King St (33,043 m²), Collins St, Swanston St to Elizabeth St (30,005 m²) and Collins St, Exhibition St to Spring St (18,864 m²), each with apartments rising in the same block. Further north the gain is far bigger than the loss: La Trobe St, Victoria St to Mackenzie St lost 12,420 m² of office and gained 102,021 m² of homes, and Lonsdale St, Elizabeth St to Queen St gained 86,528 m².
These are blocks, not buildings. The census counts what each block holds; office falling and apartments rising in one block can be one building converted, or an office demolished and an apartment tower built beside another. Across the CBD, office fell in 45 blocks and grew in 25 where both years are published.
How it was made in Plotlay
The census is published as a table keyed by block_id, a row for each block and year. The block outlines are a separate file, so the table can’t be mapped or measured on its own. One Plotlay project joined them:
- A map on the Geoscience Australia base map and a data layer called Blocks.
- Pick from a layer, Take all. The council’s CLUE blocks file: all 603 blocks in one step, keyed by their block_id.
- Link from an address. The census table straight from the council’s portal (office, apartments, hotels, student housing and all floor space), keyed by
block_id, with the census year as the time axis. A second table names each block from the streets round it. - Formula columns: each use in 2002 and 2024 (
sum(office, year = 2024)) and the change. A figure the census holds back stays empty, so no change is worked out for it. - Colour by apartments as a share of the block’s floor space, through the years on the slider; then by the change in office.
- Show in 3D: each block raised by its floor space per m² of block in 2024 (how built up it is), not by building height.
- Charts and map pictures in the Analysis workspace, exported as PNG, and the layer exported as GeoJSON, TopoJSON and CSV with WKT.
Plotlay’s office figures for 2002 and 2024 match the census table block for block (largest difference 0 m²). Nothing here comes from OpenStreetMap, and no imagery was traced.
More about Data layers · Make a data layer of your own · Where the canopy runs out
Power BI
The same blocks and census table as a Power BI Project (.pbip): the CBD by year, a Shape Map of the blocks from the TopoJSON with a year slicer, the blocks that switched, and Azure Maps. It reads the data straight from the GitHub repository, so it refreshes with no gateway.
The Power BI project is in the plotlay-melbourne-cbd-use repository on GitHub, with steps to open it in Power BI Desktop.
Python
A narrated Jupyter notebook recomputes the CBD totals and each block’s change from the census table, and checks Plotlay’s formula columns against it.
Download the notebook (with its outputs) and the data. It is also in the plotlay-melbourne-cbd-use repository on GitHub.
Read it carefully
- Blocks, not buildings. The census does not say which building changed use.
- The census leaves a cell blank where it would describe two businesses or fewer. Those cells stay blank here: 13% of the CBD’s floor space in 2024 (23% in 2002) sits in held back cells, counted in each block’s total but in no use. The CBD totals above are the published cells.
- A block’s change is worked out only where both years are published for office and apartments. Hotels and student rooms add to it where they are published.
- For about half the blocks the 2023 figures are the same as 2022’s.
- The council re-aligned some block boundaries in 2025; the census figures are for the blocks as published.
Sources and licences
Floor space per space use for blocks (Census of Land Use and Employment, 2002 to 2024), Blocks for CLUE and Road Segment: City of Melbourne Open Data, data.melbourne.vic.gov.au, licensed under CC BY 4.0. Changes: five of the floor space columns kept; each block named from the streets round it. Base map: © Commonwealth of Australia (Geoscience Australia), CC BY 4.0.
Our shapes, tables and charts are CC BY 4.0 and the code is MIT: see the licence files in the data download and the plotlay-melbourne-cbd-use repository on GitHub. Map: Leaflet. Charts and map pictures made in Plotlay; the animation with Matplotlib. This is independent work using open data, not endorsed by the City of Melbourne.
Put a table on the map
Plotlay is free to start: every tool, three cloud projects.
Create free accountThen open Data layers in any map project