AI analysis · Sale forecast
What would your home sell for today?
Pick the address and the flat, then mark its condition. The forecast works the price out from flats sold nearby recently. You see the result straight away, and I don't ask for your email.
Your home.
The forecast works for flats and houses in Tartu county. From the address I read the floor area and year built from the building register, and for a house the plot size from the cadastre. Check them and correct anything that's off.
- The likely sale price
- A range that catches about 8 in 10 sales
- What pushes the price up or down
- The price trend in your area
How it works.
Address.
The address comes from the address register, the flat's floor area, floor and year built from the building register. You add the condition and correct anything the register gets wrong.
Comparison.
Code finds flats sold nearby in recent months and adjusts their prices for your flat's floor area, floor, building era and condition. Older sales are brought up to today's price level.
Range.
You get the likely price and a range that catches about 8 in 10 sales. If there are too few sales, the forecast says so plainly and gives no number.
What the forecast can't see.
Renovation and layout
The forecast only knows the condition you pick. It can't see the quality of the work, the layout, the view or the balcony.
Data
The Land and Spatial Board's sales database and the building register. Only real sales, not asking prices from listings.
Privacy
The result contains no single sale, address or flat number from the data, only aggregates. I use the address you enter only to work out the forecast.
Not a formal valuation
Automated market overview, not an expert valuation under EVS 875. A valuation report a bank accepts is made by an Arco Vara appraiser.
Questions and answers.
Flat sales in Tartu county from the Land and Spatial Board's sales database, plus building register data: floor area, floor, year built, wall material and energy class. The forecast doesn't use asking prices from listings.
I tested the model on past sales: I worked out each price as if the sale hadn't happened yet and compared it with the real price. The median error was about 9%, and the range caught about 8 in 10 sales. So one sale in five lands outside the range, more often when the flat is unusual.
Two similar flats in the same building can sell several thousand euros apart. The range shows honestly how much prices move in your area. The likely price is the one half of similar sales land below and half above.
Tartu county has enough sales, and it's the market I know best. For houses I take the living area from the building register and the plot from the cadastre, because the sales data has neither. For older houses in Tartu city and areas with few sales I don't show a number: I value those more accurately on site.
The forecast shows the direction. I'll pin the price down on site.
I'll come over, look the home over with a builder's eye and tell you what it would sell for today.