SHIRUBE

Real estate AI data services

SHIRUBE provides real estate companies and brokerages with a valuation API, asking prices per m² by building and by area, and listing trend data.

Use the data for valuations, acquisition decisions and client proposals.

Inquire about data services

Median asking price by ward (in ¥10,000 per m²)
Ward	Median
	277.5
	274.7
	227.3
	212.5
	185.4
	178.9
	170.4
	168.1
	147.1
	145.1
	138.1
	130.4
	120.7
	117.5
	117.4
	111.7
	104.7
	94.9
	89.0
	82.3
	81.0
	71.8
	71.6

Median asking price by ward (in ¥10,000 per m²)
Ward	Median
Minato	277.5
Chiyoda	274.7
Shibuya	227.3
Chuo	212.5
Shinagawa	185.4
Meguro	178.9
Bunkyo	170.4
Shinjuku	168.1
Taito	147.1
Toshima	145.1
Koto	138.1
Setagaya	130.4
Sumida	120.7
Nakano	117.5
Suginami	117.4
Arakawa	111.7
Kita	104.7
Ota	94.9
Nerima	89.0
Edogawa	82.3
Itabashi	81.0
Adachi	71.8
Katsushika	71.6

Length shows each ward’s median asking price per m². Each dot marks a building location. Buildings at the same location share one dot. Wards with a median are ordered from highest to lowest median asking price per m², followed by wards with insufficient data.

Source: Chitaku. Medians as of September 29, 2026; locations as of September 30, 2026. Some location data © OpenStreetMap contributors

Who we serve

We serve real estate companies and brokerages that use data for valuations, acquisitions and client proposals on existing condominiums.

What we provide

Valuation API

Retrieve valuation results through an API.

Use them to support pricing decisions when you value a property.

Building-level asking prices

We provide asking price statistics per m² for existing condominiums, grouped by building.

Review the asking price level of a building you are considering and use the data in client proposals.

Area-level asking prices

We provide asking price statistics per m² by area.

Compare local price levels to inform acquisition decisions.

New listings and price changes

We provide data on new listings and changes in asking prices.

Use weekly trends to assess prices and explain market conditions to clients.

Coverage and delivery

We cover eight markets: Tokyo’s 23 wards, Kanagawa, Chiba, Osaka, Kyoto, Aichi, Fukuoka and Hokkaido.

Data is updated weekly.

Data is available through an API or as data files.

Source: Chitaku. Locations as of September 30, 2026. Some location data © OpenStreetMap contributors

Each dot marks a building location. Buildings at the same location share one dot.

Examples of data fields
Reporting week	The week covered by the statistics
Aggregation level	By building or by area
Building or area	The building or area covered by the statistics
Asking price per m²	Asking price per m² statistics
New listing trends	Changes in newly listed properties
Price change trends	Changes in asking prices

The technology behind SHIRUBE data

SHIRUBE runs on the same technology as the Chitaku real estate data platform. The technologies below support the data and its uses.

Data infrastructure

Chitaku Data Engine ZDE

Records a full snapshot of every tracked listing each week. Comparing snapshots shows new listings and price changes.

Structured data cleansing
Field mapping
Listing lifecycle identification

Weekly listing records in Tokyo’s 23 wards (six weeks)

Weekly listing records in Tokyo’s 23 wards (six weeks) the third week of August 2026 to the third week of September 2026. The chart shows listing records observed each week. Source: Chitaku
Week	Listing records
Week 1 (third week of August)	28,929
Week 2	28,862
Week 3	28,949
Week 4	29,356
Week 5	29,901
Week 6 (third week of September)	30,508

the third week of August 2026 to the third week of September 2026. The chart shows listing records observed each week. Source: Chitaku

Building record matching

Z-Entity Resolution

Merges records that name the same building differently into a single building record, so listings can be viewed building by building.

Large language models
Semantic similarity
Graph clustering
String similarity
Guard rules
Unit-level deduplication

Valuation model

Returns an estimated price per m² with a prediction interval, so each estimate comes with a range.

Machine learning
Prediction intervals
Building-level Bayesian shrinkage
Price basis calibration

Price deviation model

PDM

Tracks changes over time in the gap between asking prices and a reference price level.

Spatiotemporal asking price index
Robust statistics

Market trend analysis

Measures time on market and compares how building and area prices move against the wider market.

Survival analysis
Log return regression

Location and area analysis

Grades how precisely each location is known, standardizes address formats and tests whether neighboring areas have similar price levels.

Location precision grading
Address normalization
Spatial autocorrelation

Quality control: Accuracy measured by sample inspection

Getting started

1
Contact us
Tell us how you intend to use the data and which areas interest you.
2
Confirm your requirements
We confirm the data you need and the delivery format.
3
Evaluate the data
Try a sample dataset or a trial to review the data provided.
4
Sign the agreement
Review the data scope and service terms, then sign.

Frequently asked questions

What can I use the data for?

You can use it for valuations of existing condominiums, acquisition decisions and client proposals.

Which areas do you cover?

We cover eight markets: Tokyo’s 23 wards, Kanagawa, Chiba, Osaka, Kyoto, Aichi, Fukuoka and Hokkaido.

How often is the data updated?

Data is updated weekly.

Can individuals use the service?

SHIRUBE is a service for real estate companies and brokerages. Individuals can view listing activity for each building on Chitaku free of charge.

Visit chitaku.ai

Data service inquiries

Contact us about data for valuations, acquisition decisions or client proposals. We will review your inquiry and reply within 24 hours.