Our property price methodology

Learn how RealAdvisor builds clear, reliable and monthly updated property price estimates for every area.

How are our property prices calculated?

Carlos López Roa
12.02.2026
4 min
Table of Contents

Overview

At RealAdvisor, our goal is to provide reliable and transparent property price estimates for every street, neighborhood, and city we cover.
Our approach combines verified data, spatial modeling, and continuous updates to reflect the real estate market as it truly evolves.

Rather than relying on a single source or formula, our models bring together thousands of market observations, location insights, and statistical analysis to estimate the fair price per square meter, both at the property and area level.


1. Data Foundations

Every estimate begins with high-quality data. In Switzerland, our models rely on one of the largest and cleanest real-estate datasets available, combining millions of market observations.

We use several verified and complementary sources to ensure accuracy and representativeness:

  • Listings data

Actual properties currently on the market, including detailed characteristics (type, condition, location, features).
In Switzerland, this represents over five million listings processed over the last five years, resulting in 463,784 curated sale listings and 2,168,859 curated rental listings after quality filtering.

  • Professional data

Valuations and appraisals conducted by certified real estate experts, providing reliable ground-truth estimates where public data may be limited.
Our Swiss models incorporate 433,445 expert valuations reviewed and validated by professionals.

  • Geospatial data

Information about streets, neighborhoods, buildings, and surrounding natural or urban features that influence property values (e.g., proximity to parks, lakefronts, transport).
Our Swiss geospatial base includes 195,598 streets and 2,459,085 buildings, each enriched with detailed locational attributes.

  • Official or structured reference data

Authoritative data sources used to contextualize market dynamics, assess price evolution, and validate trends, including government datasets, cadastral maps, and reference geospatial models such as:

All data is cleaned, standardized and verified before use, ensuring that our models are trained only on trustworthy information.


2. Temporal Adjustment

The property market changes constantly.
A sale made two years ago doesn’t reflect today’s conditions, unless we adjust it.

That’s why every transaction we use is normalized to current market conditions.
Using historical price trends, we align past sales with today’s reality, allowing our models to learn from data that’s comparable in time and truly relevant to the present market.


3. Data Quality and Reliability

Before being used in our models, all data goes through a rigorous quality control process.
We carefully verify location accuracy and remove incomplete or inconsistent records.

This ensures our price estimates are based only on reliable, consistent, and geographically precise data.


4. Understanding Local Context

A property’s value depends on far more than its size or number of rooms. Location matters.
We enrich every property with detailed spatial information that helps our models capture how the environment influences prices.

For example, we account for neighborhood characteristics, street-level differences, and proximity to features like coastlines, parks, or city centers.
In areas with little available data, we use geospatial techniques to improve precision and maintain consistency across regions.


5. The Honeycomb Model

Our valuation system is based on a two-scale “honeycomb” structure, designed to balance precision and coverage.

Small Honeycombs

These fine-grained cells estimate prices for individual buildings and parcels, allowing us to represent local variations at the street or block level.

Large Honeycombs

These larger cells summarize prices for neighborhoods, municipalities, and larger regions. They also capture more detailed statistics, such as price distributions and percentiles by property types and number of rooms.

This two-tier system ensures both local accuracy and regional insight and readability.


6. Street-Level Prices

Between individual properties and neighborhoods, streets often represent the most intuitive level of comparison.
We calculate street-level price indicators by analyzing nearby properties with similar characteristics.
These values serve as a bridge between the detailed honeycomb grids and the broader area statistics, helping refine our local models and provide a clearer picture of how prices vary from one street to another.

Like all our estimates, these street prices benefit from the same data preparation, time adjustment, and quality control processes, ensuring consistency across every scale.


7. Modeling and Validation

Each geographic area, whether a city, district, or neighborhood, is modeled independently.
Our spatial models predict prices per square meter across the honeycomb grid, creating smooth, realistic, and interpretable maps of market values.

Before publication, every model is thoroughly validated through automated tests and expert review.
We cross-check results between neighboring areas and manually review any regions that show unusual patterns.
The result is a consistent, dependable view of the market.


Beyond static prices, we also provide price evolution over time.
To do this, we use an in-house model that estimates how prices change month by month, even in areas with limited recent data.
This makes it possible to follow long-term market trends for every location.


9. Continuous Updates

Real estate markets change constantly, and our data keeps up.
Our models are updated every month with the latest listings and transactions, ensuring that our price estimates always reflect the most recent market conditions.
This regular refresh keeps our maps and statistics accurate, relevant, and in line with today’s property trends.

Key Facts

RealAdvisor’s pricing engine is built on a robust foundation designed to deliver trustworthy, precise and always up-to-date property values.
It relies on:

  • A curated, high-quality Swiss dataset combining listings, expert valuations and official sources.
  • Comprehensive geospatial coverage, enabling building- and street-level precision across the country.
  • Time-adjusted historical data to reflect current market conditions with accuracy.
  • Advanced spatial modelling, including our proprietary two-scale honeycomb system.
  • Rigorous data quality controls to ensure reliability and consistency.
  • Monthly updates incorporating the latest market signals and ongoing validation.

This combination of verified data, advanced modeling and transparency allows RealAdvisor to provide accurate property prices whether you’re analysing a street, a neighborhood or an entire city.

FAQ

What makes our pricing model trustworthy?

Our model is built on a curated, high-quality dataset that combines listings, expert valuations, geospatial information and historical trends. All data goes through strict validation and quality-control processes before being used, ensuring that our estimates are based only on reliable, consistent and geographically precise information.

Who oversees and reviews our price models?

Our pricing models are reviewed by real-estate analysts and valuation experts who verify the consistency, accuracy and realism of the results. This expert oversight ensures that any unusual patterns are identified and corrected before publication.

How often are our price estimates updated?

We update our estimates every month, incorporating the latest listings, expert valuations and market shifts. This ensures that our prices reflect the most current conditions across Swiss cities, towns and neighborhoods.
Carlos López Roa
Carlos is an experienced data scientist and machine-learning engineer with over eight years of expertise in designing, building and operating scalable data platforms and advanced predictive models. At RealAdvisor, he oversees the technical review of our pricing methodology, ensuring that every model is rigorously validated, statistically sound and aligned with industry best practices. His deep experience in production-grade ML strengthens the reliability and consistency of our valuation system.