ChatHome publishes aggregate statistics computed from the property adverts it indexes and from its own travel-time model, as open datasets. Each dataset has its own page, licence, CSV and Parquet files, a README and a machine-readable datapackage.json.
Current release 2026.10, generated . Every file of the release, with its size and SHA-256 checksum, is listed in manifest.json.
Median advertised sale and rent prices per m² for Luxembourg's 100 communes and Luxembourg City's 24 quarters, with a monthly series and the official commune registry.
How advertised homes for sale and rent are spread across energy-performance classes, with asking prices per m² by class, in Luxembourg and the neighbouring border areas of France, Belgium (Wallonia) and Germany.
Modelled driving, public-transport, cycling and walking times from Luxembourg's 100 communes and the main French and German cross-border home towns to six Luxembourg job hubs.
Release 2026.10-5 packages source release 2026.10, assembled 2026-10-18. Each table keeps its own observation dates. These are dated asking-price aggregates and modelled journeys.
Luxembourg housing asking prices by commune and month
Asking prices per square metre of living area for apartments and houses offered for sale or rent in each of Luxembourg's 100 communes and in the 24 official quarters of Luxembourg City. Values are the median and the 25th and 75th percentiles of the residential listings that were active on ChatHome when the statistics were computed; ChatHome collects these listings from estate-agency websites. The data has four segments (sale apartment, sale house, rent apartment, rent house), a current snapshot per commune and per quarter, and a monthly history in long format that grows with each release. A median or quartile is published only when at least 5 listings support it; otherwise the cell is empty, never zero. With few listings, a median or quartile can equal a single advert's asking price per m²; the adverts are public, so this reveals no private data. For reference, the commune snapshot also carries, where available, two official Observatoire de l'Habitat series from data.public.lu: the mean asking rent over a rolling 12-month window and the mean registered apartment sale price. These differ from the ChatHome figures in statistic (mean, not median), window, source and coverage, and the registered price is a transaction figure while every other price here is an asking price. Each commune row carries the official 4-digit commune code used by STATEC and data.public.lu.
Cite: ChatHome (chathome.lu), Luxembourg housing asking prices by commune and month, release 2026.10-5. Source dates and publication thresholds apply.
Advertised homes and asking prices by energy class, Luxembourg and border regions
How the homes advertised for sale and for rent in Luxembourg and in the neighbouring areas of France, Belgium (Wallonia) and Germany are distributed across energy-performance classes, and what they ask per square metre in each class. Each country keeps its own certificate scheme and headline indicator: the Luxembourg CPE primary-energy class (A+ to I), the French DPE energy-consumption class (A to G), the Walloon PEB class (A++ to G) and the German Energieausweis class (A+ to H). Letters are never pooled or converted across schemes. A listing whose class is missing, on another axis or outside its scheme's scale counts as not stated, and not-stated rows are published so the gap stays visible. The source is a one-day snapshot of the public ChatHome catalogue of active listings, which ChatHome collects from estate-agency websites. Counts and shares need at least 5 listings, because smaller counts are unreliable; the adverts are public and this is not a confidentiality guarantee, as a hidden count can sometimes be worked out from the shares. Median, 25th and 75th percentile asking prices per m² and new-build shares need at least 10. A second table gives the class mix by Luxembourg canton. The shares describe advertised stock on the snapshot date. They are not the national certificate registers and not the housing stock.
420 rows · CC-BY-4.0 · 2026-10-18/2026-10-18
Modelled travel times from Luxembourg and cross-border communes to six Luxembourg job hubs
Modelled door-to-door travel times from every Luxembourg commune and from the French and German communes where at least 100 residents work in Luxembourg (IGSS counts, at most 400 communes) to six Luxembourg destinations: Luxembourg station, Kirchberg, Cloche d'Or, the city centre, Belval and Findel airport. Each row gives one origin, destination and weekday scenario: driving off-peak and at the morning and evening peaks, public transport in the morning and evening peaks, walking and cycling. Road, walking and cycling times come from the open-source Valhalla router on OpenStreetMap; public-transport times come from the R5 router on official GTFS timetables, with the 80th percentile published next to the median. Peak driving times are modelled with a national congestion ratio, not measured traffic. A median over 120 minutes is reported as no journey found, never as zero.Every row states when its timetable or road graph stops being valid. Belgian communes are not included while the licence of their centroid source is unresolved. Because the times are derived from OpenStreetMap, the dataset is published under the ODbL.
Published preview: commune_hub_travel_times. Ordered by commune or dwelling/class identifier, not an investment score. Missing values are shown as —.
country
national code
hub id
mode
scenario id
minutes p50
status
LU
0101
kirchberg
Luxembourg Commune Asking Prices and Modelled Commutes
Observed commune asking-price differences versus extra modelled morning commute minutes to Kirchberg, Cloche d'Or, Gare and Belval, relative to Luxembourg City. Luxembourg only; no matched cross-border prices are available.
Published preview: commute_price_arbitrage. Ordered by commune or dwelling/class identifier, not an investment score. Missing values are shown as —.
commune code
commune name en
property type
hub id
mode
price sample
reference price sample
Luxembourg CPE Asking-Price Differences by Energy Class
National advertised price medians compared with class A within each tenure and property type. This descriptive comparison preserves source suppression and does not estimate thermal-class effects, commune premiums or renovation eligibility.
Published preview: cpe_class_price_comparison. Ordered by commune or dwelling/class identifier, not an investment score. Missing values are shown as —.
2026-10-18, release 2026.10: commune_rent_price_quarterly is withdrawn. Release 2026.08 labelled it as Observatoire de l'Habitat advertised rents, but it came from ChatHome's retired rental_price_index: it carried size brackets, medians and quartiles that the Observatoire does not publish, and 4,428 of its 9,470 rows had fewer than 5 observations.
2026-10-18, release 2026.10: commune_sale_price_quarterly and commune_asking_price_annual are no longer published. They copied official Observatoire de l'Habitat series that could not be rebuilt reproducibly: the history is no longer stored and the stored commune codes were not the official ones. Use the originals on data.public.lu (CC0).
2026-10-18, release 2026.10: frontalier_belt_communes (2,321 French communes) and cross_border_source_status are withdrawn. travel_time_localities in the travel-times dataset covers only the French and German travel-time reference communes, with official codes, names and IGSS cross-border worker counts, and carries no population or postal codes.
2026-10-18, release 2026.10: Release 2026.08 published 25th and 75th percentiles in commune_prices_latest for 57 cells (in 50 communes) with only 2 to 4 observations. Quartiles are now empty below 5 observations.
2026-10-18, release 2026.10: commune_name_lu in commune_registry was empty in release 2026.08. It is now filled.
2026-10-18, release 2026.10: Release 2026.08 said the data was regenerated monthly. No later release was produced before this one; releases are now run by hand and each one states its date.
What this data is not
Asking prices are not transaction prices. They are what sellers and landlords advertise, and they describe the homes on the market, not the housing stock.
Travel times are modelled, not measured. A journey not found within 120 minutes is stated as such, never as a zero.
An empty cell is never a zero. Medians and quartiles are published only for cells with at least 5 adverts (10 for prices by energy class); below that the value is empty.
No personal or listing-level data. Every row is an aggregate, a modelled travel time or an official registry entry: no listing identifiers, descriptions, photos, addresses, agency identities or user records.
Licence, citation and releases
The two price datasets are released under CC BY 4.0. The travel times are computed on OpenStreetMap and are released under the ODbL 1.0, so a database built from them stays under the ODbL. Each dataset page gives the exact citation and the attribution its sources require.
Releases are run by hand and each one states its date. The files of a release never change: a correction is a new release. Versioned file links are permanent, and the latest links always serve the current release.
Published preview: energy_class_asking_prices. Ordered by commune or dwelling/class identifier, not an investment score. Missing values are shown as —.
Cite: ChatHome (chathome.lu), Advertised homes and asking prices by energy class, Luxembourg and border regions, release 2026.10-5. Source dates and publication thresholds apply.
Cite: ChatHome (chathome.lu), Modelled travel times from Luxembourg and cross-border communes to six Luxembourg job hubs, release 2026.10-5. Source dates and publication thresholds apply.