The same price does not tell the same story. In 2025, houses sold in Werribee and Manor Lakes had the same reported suburb median: $650,000. Yet their annual medians followed different paths over the preceding decade. This project asks what a current price band can reveal about the past—and what it cannot tell a prospective buyer about genuine affordability.
Research question: Among metropolitan Melbourne localities with 2025 median house sale prices between $600,000 and $700,000, how different were their 2015–2025 price histories, and what important information is missing from the headline median?
The intended audience is prospective Melbourne house buyers—especially readers comparing relatively lower-priced suburban areas. A suburb median is a useful starting point, but it summarises houses sold in a particular year, not the cost of every property for sale. Because the data cover house prices rather than household incomes or borrowing costs, this article uses the precise term “lower-priced” instead of claiming that these suburbs are affordable for everyone.
The analysis is deliberately limited to 42 localities meeting the specified 2025 price band. Their reported 2025 median is $655,000. The history that follows holds this same group of places fixed; it does not show which suburbs were the cheapest across Melbourne in each earlier year.
| Locality | 2015 median | 2025 median | Change in annual median |
|---|---|---|---|
| Manor Lakes | $385,000 | $650,000 | 68.8% |
| Werribee | $335,000 | $650,000 | 94.0% |
Werribee’s reported median increased from $335,000 in 2015 to $650,000 in 2025 (+94.0%); Manor Lakes rose from $385,000 to $650,000 (+68.8%). The difference is 25.2 percentage points of median-price change. It is not a comparison of returns on identical houses, and it does not identify which suburb is a better future purchase.
The following interactive figures widen the lens from this two-suburb contrast to the full selected set, then check whether missing or flagged values affect the comparisons.
Primary source: Valuer-General Victoria, Houses by suburb 2015–2025, in the Victorian Property Sales Report series. The government explains that annual medians summarise sales recorded in a locality during a calendar year; they are not repeat-sale measures of the same properties. Official annual source.
Inclusion rules: The workbook was filtered to the working 31-council metropolitan Melbourne definition; qualifying localities have a numeric 2025 house median in the chosen inclusive range; their 2015–2025 data were retained; preliminary 2026 values were excluded. Region labels (‘West’, ‘North’, etc.) are working grouping labels, not government-defined statistical regions.
Geographic caution: East Warburton was added using an official Yarra Ranges locality map. Beveridge, Woodstock, Hastings West and Merinda Park remain out of the main sample pending more geographic verification. Council-name screening is not a complete geographic-information-system polygon audit. See the research data-methodology notes.
| Metric | Value |
|---|---|
| 2025-selected localities | 42 |
| Annual locality-year records | 462 |
| Missing annual median values | 16 |
| Low-sales flagged locality-years | 26 |
| Localities missing a 2015 value | 5 |
| 2015–2025 comparisons with usable endpoints | 37 |
Hover to see the council, region, exact median and low-sales marker. Use the legend to isolate regions. A dot rather than a bar avoids implying comparisons start at zero.
This stacked histogram divides the $600k–$700k band into $10,000 intervals. Hover to inspect counts. The top price of $700,000 is included in the final interval.
Each line plots the unweighted median of available suburb-level annual medians within that working region. Small-sales flagged yearly observations are excluded from this summary. Hover points to see how many localities contributed that year; a line is not the transaction-weighted metropolitan median.
Use this heatmap to locate missing annual prices and unusual changes without interpreting blank cells as zero. Hover to inspect a particular locality/year and its small-sales marker.
Each point is a locality with a numeric 2015 median. Percent change compares different annual sale medians, not the realised gain from reselling the same house. Diamond-shaped markers identify localities with a low-sales flag at either endpoint; use particular caution for very large percentage changes.
First, the 2025 price window hides different starting points. Werribee and Manor Lakes reached the same $650,000 reported median in 2025, yet their 2015 medians differed by $50,000. Their respective changes of 94.0% and 68.8% demonstrate why a current suburb median cannot stand in for its historical path. They also show that rising percentages partly reflect different bases.
Second, the pattern is not confined to two suburbs. The selected group includes 42 metropolitan localities; 21 are in the working West grouping. Among the 32 localities with two available, unflagged endpoints, reported annual medians increased by 49.0%–129.8% over the decade. These are changes in annual suburb medians, not returns from selling the same home ten years later.
Third, uncertainty is visible rather than hidden. The dataset contains 16 missing annual prices, 26 low-sales flagged locality-years, and 5 localities without a numeric 2015 baseline. Figure 4 makes these limitations inspectable, while Figure 5 marks low-sales endpoint comparisons. For example, Kalkallo’s +251.7% headline change is based on a flagged endpoint: it is not suitable as a representative result.
A $650,000 suburb median is a starting point for questions, not an answer about value. Buyers comparing places in the same 2025 price band should examine longer-run records and data quality, then investigate factors this dataset does not contain: household income, dwelling type and size, transport access, local amenities, financing costs and current listings. The supplied data cannot establish true household affordability, predict future prices, or rank suburbs as investments.
A further limitation is deliberate selection on the final year. Since these 42 places were chosen using their 2025 prices, the project describes their histories. It cannot answer how many suburbs in the whole city were in this price bracket during 2015, or whether Melbourne became more affordable overall. Answering those broader questions needs a different design and further datasets.
The 2025 price threshold was chosen before viewing historical
outcomes. Therefore the past-year comparisons describe the
present-day selected cohort; they cannot show how the population of
lower-priced Melbourne suburbs changed year by year. House sale medians
can shift partly because the mix of properties sold differs from year to
year. Localities with few sales have more volatile annual medians;
^ markers are retained for transparency. Missing prices are
not imputed.
For the council boundary decision, this report uses the Victorian Parliamentary Budget Office’s 31-council metropolitan grouping. This is not the same as a Census Greater Capital City Statistical Area, an urban-growth boundary, or a full geospatial assessment of exactly where individual sales occurred.
| Locality | 2025 median (AUD) | Potential council | Decision |
|---|---|---|---|
| BEVERIDGE | 650000 | Whittlesea / Mitchell | Keep excluded pending split-geography sales evidence |
| WOODSTOCK | 647000 | Whittlesea / Mitchell precinct | Keep excluded pending geography check |
| HASTINGS WEST | 665000 | Mornington Peninsula | Keep excluded pending reporting-area boundary evidence |
| MERINDA PARK | 671000 | Casey | Keep excluded pending reporting-area boundary evidence |
Source audit: The working dataset contains 16 unavailable annual medians, 26 low-sales flags, and 5 localities without a 2015 median. The source workbook can be checked below when distributed alongside this report’s R code.
## **Original-workbook numerical verification:** 42 localities checked; 0 unmatched 2025 values.Department of Transport and Planning, Victoria. (2026). Victorian Property Sales Report — Median House by Suburb Time Series. DataVic. https://discover.data.vic.gov.au/dataset/victorian-property-sales-report-median-house-by-suburb-time-series (CC BY 4.0).
Valuer-General Victoria. (2026). Property sales statistics: Houses by suburb 2015–2025. Victorian Government. https://www.land.vic.gov.au/valuations/resources-and-reports/property-sales-statistics
Parliamentary Budget Office, Victoria. (2025). Regional and metropolitan investment 2025–26. https://static.pbo.vic.gov.au/files/PBO_Regional-and-metropolitan-investment-2025-26_PUBLIC.pdf
Department of Transport and Planning, Victoria. (2019). Yarra Ranges locality boundary map. https://www.land.vic.gov.au/__data/assets/pdf_file/0034/499831/Yarra-Ranges_V51_Mar2019.pdf
RMIT University Library. (2026). Acknowledging the use of AI tools. https://rmit.libguides.com/referencing_AI_tools/acknowledging
Generative AI disclosure. OpenAI ChatGPT (GPT-6; accessed 22 October 2026) contributed substantively to developing this project, including assistance with project structure, dataset filtering and geographic-screening decisions, R Markdown and interactive Plotly code, source-data validation and checks, correction of chart legends and symbol mappings, explanatory captions, and drafting and editing the main narrative and findings. The topic (Melbourne suburban house prices) and $600,000–$700,000 focus were chosen by the student. Figures are generated using R code and based on official Victorian Government data; claims and code should be independently checked before submission. This acknowledgement is additional to, not a substitute for, the course’s separate AI Declaration form.
OpenAI. (2026). ChatGPT (GPT-6). https://chatgpt.com
This report uses five interactive plotly visualisations
created through R code with the uploaded
government-derived CSV. The HTML copy distributed with this package was
adapted from the successfully knitted R output provided by the student;
its text and widget legend fixes were adjusted to match this final Rmd.
Hover, interactive legends and Plotly’s zoom controls are available in
the knitted HTML. Exporting static screenshots would remove the
interactivity.
## R version: R version 4.6.0 (2026-04-24) | Localities: 42 | Annual rows: 462