What can be recreated from public data?
National and local completed-sale trends, a transparent purchasing-power index, economic context, official construction insolvency events, school inspection summaries and a sourced supermarket-location catalogue.
TwentyCi describes a much broader longitudinal property and homemover dataset. Public completion records cannot reproduce proprietary listing events, sale fall-through rates, time to sell, mover identities, its AVM or a complete mortgage-product catalogue.
TwentyCi’s published scopeHow resale £/sq ft will be calculated
Match standard existing-home transactions to an appropriate EPC using an authorised UPRN route. Use reliable floor area and a certificate valid for the sale period. Calculate each transaction’s price divided by floor area × 10.7639104167, then take the median for the stated cohort.
Exclude ambiguous matches and invalid areas; disclose matched coverage, sample size and the unmatched share. EPC access requires a One Login bearer token. Non-address fields and address/postcode reuse have different licensing conditions.
No £/sq ft values are currently published. An account connection and validated match sample are still required.
School and supermarket association studies
Compare similar completed sales using floor area, dwelling type, age, tenure, sale quarter and fine-grained location controls. Show sample counts, overlap, unmatched properties and uncertainty for each estimate.
School admissions boundaries are not a circle around a school. Inspection grades must use the framework and date relevant to the transaction. New domain grades should not be converted into old overall grades.
Waitrose and Lidl choose different trading locations. A raw nearby-price difference can reflect local incomes and property mix. Store-opening studies also need real opening dates; a hygiene-rating date is not an opening date.
Daily checks; honest revisions
The cloud refresh checks committed primary sources, validates their release dates and updates reviewed rows. A daily run may correctly leave a monthly or annual statistic unchanged. Source failures retain the previous successful observations.
HPI, Price Paid and earnings data can revise. Historical figures belong to the latest retrieved release, not an invented real-time feed. Registered volumes remain incomplete for recent transaction periods.
Newsletter delivery uses confirmed consent, a verified sender and persistent delivery receipts. Provider credentials stay server-side; subscriber addresses are separate from the public data snapshot.