British Film Database (our own project) in detail

How we did it

A reference work goes stale the day it is published unless someone keeps it current. The site is built from open data. The films and most of their details come from Wikidata, which places its data in the public domain. Cast billing and some studio names come from Wikipedia. Photographs come from Wikimedia Commons, each credited. Where a film can be streamed or bought comes from JustWatch through TMDB, checked weekly. Every source is named on the page that uses it.

A scheduled build fetches the sources, reconciles them and regenerates about 34,000 pages: films, people, decades, genres, studios and certificates. Nobody edits a film page by hand. The features and the about page have a visual editor in TinaCMS. Search, a watch list and the where-to-watch panel run in the browser from the built data.

One run of Google's hosted PageSpeed Insights on 28 October gave 1.2 seconds and 100; another on the morning of 30 October gave 2.4 seconds and 98. Later that day, after the site was changed, five Lighthouse runs gave a median of 1.7 seconds. A spread like that is why we publish medians and the range, not one run. For a manufacturer the same engineering is a product library that updates from the data you already hold, so the data sheets, the finder and the search never disagree.

Results

• Main content on screen, mobileResult1.7 s
• Performance score, mobileResult100
• PagesResult34,057
• Films listedResultMore than 9,500
• RefreshResultWeekly, automatic

Back to the headline changes