Data first. Filter, sort, export.

For each neighborhood block (chōchō, the 12-digit code), we compare its price-to-rent against a regional baseline — the same prefecture and population-density band. A positive z-score means the block is priced higher than its peer set; a negative z-score means lower. The magnitude is objective and computed from observed data. We do not read direction as opportunity; the investor forms their own thesis. A regression-model toggle (v1 baseline / v2 adding development-plan proximity) lets you see how the estimate shifts.

An estimated yield is derived from observed rent and price for the block. It is a gross estimate: taxes, expenses, vacancy, and repair costs are not included. It exists to let you sanity-check a block against its neighbours on the same basis, not to project a return.

We surface proximity to 164 confirmed development plans, sourced from MLIT and operator and municipal disclosures. Proximity is shown as context — location, type, status, and source — not added to a score as points. Peer-reviewed evidence shows most of a plan's price effect is priced in at announcement, and the sign can reverse between thick city markets and thin regional ones; uniformly scoring "near = plus" would mislead. See the methodology for the full reasoning.

Preset filter sets can be saved and surfaced as alerts when new block-level signals cross your thresholds. Results export to Excel/CSV for your own modelling, and presets can be re-run for year-over-year comparison. These convenience features sit behind the paid tier; browsing the underlying figures stays free.