Завантаження інструмента…
REALTYHUB · PRICE TRENDS
Follow property prices over time
Choose your segment and up to eight years of history to follow recorded amounts and explore an experimental price outlook where enough data is available.
Choose up to four neighbourhoods to compare their price histories, then update the view.
Area boundaries · Drag to move · Scroll to zoom
Loading your trends…
Your filters have changed. Update the view to see the new results.
The count and charts cover your selected history. Median and average use its final 12 complete months to show the price level near the end of that history. The percentage compares these with the preceding 12 months, so both windows include the same calendar months.
Recorded price trends
See whether typical recorded amounts rose or fell over your chosen years.
Update the selected history to view its forecast.
Change between two points
Choose two dates to see how much median and average amounts changed, in percent.
Different properties can be recorded at each date, so this change does not tell you how much one property has gained or lost. Amounts exclude VAT.
Activity behind the prices
See how recorded activity changed alongside the price history.
Counts use your chosen monthly, quarterly or yearly view. * means the period is only partly covered by the selected dates or available data.
LOOKING AHEAD
What prices are forecast next?
Choose a horizon to see the price outlook for your selected property type and area. Based on recorded amounts; an experimental scenario, not a guaranteed future price.
Upper and lower scenarios follow the range of past forecast errors; they are not separate predictions or guaranteed limits. The thin dotted line shows the central path. Select a quarter to compare all three values.
How the projection is checked
We first try forecasting earlier parts of your selected history, where the actual results are already known. We compare several simple methods and show a projection only when there is enough history and the past errors pass our checks.
- At least six years of suitable quarterly data are needed. Price projections also need at least ten valid amounts in every quarter.
- The line is one possible continuation. The shaded range comes from past forecast errors and has no guaranteed probability.
- A longer look-ahead is more uncertain. A 24-month projection is a scenario to explore, not a price estimate for a property.
Calculation details and checks
The chart shows the most recent two years of observations; training and validation use the eligible selected history. Select a projected point to read its amount or count.
Six years are required for 6- and 12-month scenarios; eight years for a 24-month scenario. Prices also need ten valid amounts in each complete quarter. Sparse recent quarters are not skipped or filled. Counts retain zero quarters.
For each historical test date, we choose the method using only earlier data, then check its forecast against later observations. We test every step of the requested horizon. The three simple references are the last quarter, the same quarter last year and the recent four-quarter average. A damped trend must reduce selection error by at least 10%.
At least three selection origins and three later validation origins are required. Test paths can overlap, so the checks are not independent repetitions. Errors and historical range coverage are reported by horizon. The 35% median relative-error limit is an experimental setting. Shading uses the 80th percentile of absolute log errors from later validation; it is not a calibrated probability interval. No fixed property cycle is assumed.
Method references: Forecasting: Principles and Practice — benchmarks and time series cross-validation.
How to read these figures
What these trends tell you
The charts track recorded amounts in RealtyHub’s research. They are not adjusted for inflation, size, quality or changes in the properties recorded. A rising median or average does not show how much one property has increased in value.
Two ways to read the price chart
“Each month / quarter / year” uses the amounts recorded within that period. “12 months ending at each point” combines the amounts from that month and the 11 before it. This can make the longer trend easier to see. We calculate from the individual amounts, not by averaging monthly medians.
Areas, counts and dates
Each data point is counted once in the combined selection. A record linked to several areas can appear in each area chart, with its full amount. Contract and sale records have not been linked to identify unique property sales. No matching records does not prove that no activity occurred.
Records are grouped by acceptance date. The latest 12-month cards use complete months within your selected dates. Neither the statistics nor the projection use data outside your chosen history.
Amounts and VAT
Amounts exclude VAT; where applicable, VAT is added on top. “Both” includes records regardless of the VAT flag. Price figures use valid, positive amounts. Other matching records stay in the counts. Very high and low amounts and partial-property interests are kept.
Automatic exclusions
Records flagged as loan restructuring, antiparohi, bulk transaction, other change, site value or missing price are always excluded. You choose the VAT setting.
Explore changes in the selected data, alongside the number and mix of observations behind them.