Transparency

Which sources do we use?

We combine public market statistics with open location and context data. For each source we state the role and what it can and cannot tell.

Price data

• Fednot property barometerPrimary

  Median sale prices (incl. p25/p75) at sub-municipality or municipality level, based on actual notarial deeds. This is our primary price anchor. Fednot publishes a 1–5 confidence indicator; exact transaction volume is not published.

• StatbelFallback

  Official public sale data per municipality (trailing year), with transaction counts, median and p25/p75. Fallback when Fednot is unavailable; also house types and provincial/national historical series.

Location and context data

• Geocoding (Nominatim · geocode.xyz · Geopunt)

  Address → coordinates and location type. Primarily OpenStreetMap Nominatim; then geocode.xyz; final attempt Geopunt (Flanders/Brussels). No price data.

• iRail · open GTFS

  Train stations within 2.2 km and public transport stops within 1.2 km via open station and stop feeds (NMBS, De Lijn, STIB/MIVB, TEC). No price data.

• Neighbourhood POI & green (OpenStreetMap bundle)

  Parks and green space within 1.5 km; shops, schools and sports within 1.2 km of the address from a pre-built OpenStreetMap extract (Geofabrik Belgium). Refreshed at least monthly in our build pipeline; no live POI API while generating a report. No price data.

Map and POI data: © OpenStreetMap contributors.

How we calculate the price

We always follow the same three steps. Every report shows which choices were made for that specific address.

1. 1. Source anchor

   We start from the Statbel municipality median (rolling 4-quarter window) as the price anchor. Where available we refine with Statbel statistical sectors (NIS7). Fednot serves as a secondary reference.

2. 2. Local refinement and corrections

   On top of the anchor we apply corrections based on property characteristics and location. When no hard local transaction layer is available, fine-grained location stays context - not a separate neighbourhood median.

3. 3. Guardrails

   We only show a finer level starting from a minimum of valid transactions and a reliable sector mapping. If that falls short, we deliberately stay on a broader benchmark to avoid false precision.

Corrections on the price anchor

The anchor is the starting value. Three factor categories correct it upward or downward. The report shows the exact percentage applied per factor.

• Size profile

  Surface area and how it compares to the local average. Very small or very large homes often deviate from the median. Typical effect: roughly ±5% to 10%.

• Condition & energy

  Condition, EPC score, renovation potential or energy label. A poor energy score pushes value down; a recent renovation lifts it. Typical effect: roughly ±5% to 10%.

• Micro-location

  Mobility (public transport, train, cycling), green space and nearby amenities (OSM neighbourhood bundle) and location context. Without a hard neighbourhood transaction layer this does not replace a neighbourhood median, but does nuance above or below the municipality anchor. Typical effect: roughly ±3% to 7%.

After corrections a revised core value remains. Around that we draw a band (see next section).

Band and confidence

The band shows how tight or wide the estimate is. That depends on how much input we have and how well the local data covers the address.

A normal range is around ±7%. With very solid coverage it can be tighter; with thin data we deliberately widen it.

What drives the band:

• Transaction volume in the source - more deeds means more confidence.
• Which level the anchor uses (sector, sub-municipality or municipality).
• How many property characteristics we know (the more complete, the sharper the corrections).
• Quality of the address match (address → sector/neighbourhood).

Confidence score

The score in the report (e.g. 71/100) summarises how reliable the input is for this specific estimate. Lower scores go together with wider ranges and more indicative copy.

How extra input sharpens the estimate

Every extra signal makes the corrections more concrete. Where a factor first leaned on a market-wide average, it becomes more personal - narrowing the uncertainty around your range.

• Bedrooms & living area

  Sharpens the size profile. Without this input we use a market-wide average; with the right room count the correction matches comparable homes better.

• Condition

  A direct, subjective correction. Without a condition indicator we stay neutral; with ‘renovated’ or ‘to refurbish’ the value shifts clearly.

• EPC

  Refines the energy correction: a poor EPC (E/F) pushes down; a good one (A/B) lifts up. Typically a few percent and narrows the uncertainty around the energy correction.

Result: the core value shifts closer to the right level and the band around your range is usually a bit tighter.

How to read the sources panel in your report

In your report under ‘All sources and context’ you'll find the same story applied to your address. This is what each block means:

• At the top: your confidence score, the primary source with reference year and a short headline about which benchmark we use.
• Scope: the exact level (sector, sub-municipality or municipality) where the price anchor is computed. Neighbourhoods within the same municipality can differ strongly - that nuance comes from location and neighbourhood data.
• Score context: which non-price signals (mobility, green space, location) contribute to the score, without replacing a transaction layer.
• Limitations: what we do not know, cannot refine or cannot make hard for your address. This is also where we flag a missing build year, bedroom count or EPC.
• Price anchor and corrections: the starting amount per m², the amount after corrections, the resulting range (±%) and the three factor categories (size profile, condition & energy, micro-location) with their applied percentages.
• Sources: per source we show whether it counts (primary, secondary, reference, context), its scope, its year and a short note. Sources marked ‘contributes to score’ feed the neighbourhood and location score, not the price anchor.
• Method: a short summary of the three steps (source anchor, local refinement, guardrails) as they play out for that specific report.

Limits and what we do not do

Pandwijzer is a first handhold when buying, selling or bidding. It is not a valuation, investment advice or legal advice. Even where we show neighbourhood or sector nuance, that is not a street median or an official valuation. For major decisions, speak to a notary, estate agent or bank.

See Privacy & data for how we handle your own information. Privacy & data.