ProdReady Research

How the research is produced.

The two datasets behind every figure, the checks we apply, and what the numbers can and cannot tell you.

Developer research: public GitHub profiles

  • What it is: public GitHub profiles and their public repositories, collected by our own crawler. Each finding states the date the data runs to.
  • Who counts: profiles whose stated location we can place in the UK or in mainland China. Organisation accounts are excluded.
  • Location checks: automatic geocoding misplaces many profiles. "Cambridge, MA", "Stanford, CA" and "Birmingham, AL" read as British, and "Philadelphia, PA" as Chinese. We re-check every location against its own text and drop anything that names somewhere else. A bare town name shared with a larger place abroad (Cambridge, Birmingham) only counts when a UK marker follows it, so those cities' figures are floors.
  • City populations: per-head figures use official local-authority population estimates (ONS, NRS and NISRA), named on each insight.
  • Main language: the language GitHub detects in most of a developer's public repositories, with recent ones weighted double. Markup, styling and build files do not count.
  • AI projects: repositories whose name, description or topics name a language model, model provider, agent, retrieval or vector search, MCP server, embeddings, fine-tuning or an AI coding tool. These are mentions, not proof of use.
  • Limits: this describes developers who publish on GitHub and say where they are. That group leans towards open-source contributors, academics and students. Many Chinese developers also use Gitee, so China is under-represented, and comparisons with China should be read with that in mind. Self-reported fields (employer, "available for hire") are not verified.

Employer research: UK employers' careers pages

  • What it is: the careers pages and job adverts that UK employers publish on their own websites, read by our crawler. Job boards and recruitment agencies are excluded.
  • Who counts: direct employers we have confirmed as UK businesses, with sector and region from Companies House (region is the registered office, not the job's location).
  • Employer-weighted: every employer counts once, however many adverts it posts, so a handful of large recruiters cannot swamp a figure. Sample sizes are numbers of employers.
  • Minimum samples: a figure is published only when enough employers sit behind it (30 by default). Smaller segments are withheld, not estimated.
  • Limits: employers that advertise only on job boards or through agencies are not seen. Adverts more than 180 days old are left out.

Citing and custom cuts

Quote freely with a credit to ProdReady Recruitment and a link to the insight's page. For a custom breakdown or the underlying table, contact us.

Every insight