How We Research

Our Research Process

Each neighbourhood guide goes through four stages before publication. The process typically takes 2–3 weeks per guide, depending on data availability and the complexity of the area.

Stage 1: Data Collection

We pull structured data from official government and regulatory sources. No data in a PAL guide comes from user reviews, scraped estimates, or unverified third-party databases. Our primary sources are:

• Property prices — HM Land Registry Price Paid Data. We use 12–24 months of completed transactions, broken down by property type (flats, terraced, semi-detached, detached). Prices reflect actual sold prices, not asking prices or algorithmic estimates.
• Schools — Ofsted inspection reports (via the official reports site) and the DfE’s Get Information About Schools (GIAS) database. We include every state-funded school within a reasonable catchment of the neighbourhood, filtered by Ofsted rating and distance.
• Transport — Transport for London journey planner data, National Rail timetables, and station location data. Journey times are calculated for weekday morning peak (arriving by 9am) to five key London destinations.
• Crime — data.police.uk street-level crime data, published monthly by the Metropolitan Police. We calculate a per-1,000-population risk score and compare it to the London-wide average.
• Council costs — direct from each London borough’s published council tax schedules, parking permit fees, and waste collection charges. We verify these annually against the council’s own website.
• Green space and geography — Ordnance Survey data and ONS area measurements, cross-referenced with the GLA’s ward-level housing projections for population and density figures.

Stage 2: AI-Assisted Analysis

We use AI tools to assist with data processing, cross-referencing, and initial draft generation. To be transparent about what this means in practice:

• What AI does: aggregates data from multiple sources into a structured format, calculates derived metrics (e.g., price comparisons to zone averages), identifies data gaps or inconsistencies, and generates initial editorial drafts from structured data.
• What AI does not do: make subjective judgements, write editorial verdicts, select which data to highlight or downplay, or publish anything without human review. The editorial voice, balance, and final judgement in every guide is human.

We believe this is the right balance: AI handles the heavy lifting of data processing at scale, while human editorial judgement ensures the output is accurate, balanced, and genuinely useful.

Stage 3: Human Editorial Review

Every guide is reviewed by our editorial team before publication. This review checks for:

• Data accuracy — are the numbers correct and properly sourced?
• Balance — does the guide present both strengths and considerations fairly?
• Local accuracy — does the description match the reality of the area? (We cross-check against local knowledge, recent news, and on-the-ground research.)
• Completeness — are all sections populated with current data?
• Readability — is the guide accessible to a general audience without dumbing down the data?

Stage 4: Publication and Ongoing Updates

After publication, guides are not static. We update them on the following schedule:

• Property prices — refreshed quarterly when new Land Registry data is published (typically 3–4 months in arrears).
• School ratings — updated when Ofsted publishes new inspection reports for schools in the area.
• Crime data — refreshed when data.police.uk releases new monthly statistics.
• Council costs — updated annually when councils publish new tax and fee schedules (usually April).
• Transport — updated when TfL introduces service changes, new stations, or timetable revisions.

Every guide displays a “Last updated” date so you can see exactly how current the information is.

How We Calculate PAL Scores

Each neighbourhood receives a PAL Score across six categories: Transport, Schools, Safety, Green Space, Value, and an Overall score. These are calculated from the raw data using a transparent, weighted methodology:

• Scores are normalised against London-wide averages, so a score of 5.0 represents the London median.
• The methodology is documented in full in our internal data architecture documentation.
• Scores are recalculated whenever the underlying data is refreshed.

PAL Scores are designed to be directional indicators, not absolute rankings. They help you quickly compare areas, but we always recommend reading the full guide for the nuance behind the numbers.

Corrections and Feedback

We take accuracy seriously. If you spot an error, an outdated figure, or something that doesn’t match your experience of an area, please let us know via our contact page. We investigate every report and publish corrections promptly.

We also welcome suggestions for new neighbourhoods to cover, additional data points to include, or improvements to our methodology.

Editorial Independence

PAL does not accept payment from estate agents, developers, or property companies in exchange for favourable coverage. Our guides are editorially independent. If we ever introduce affiliate links or sponsored content, it will be clearly disclosed in compliance with UK advertising standards.