UK Public Sector GitHub Observatory
Every commit carries an author name and email. Collapsing 26 per-repository author records by email gives 15 distinct identities - 15 people and 0 machines. Identity is keyed on email because it is the only stable key git offers; one person using both a work and a personal address will appear twice.
People
What stands out
1 people out of 15 account for 432 commits. The top ten alone wrote 98%. This is the usual open-source power law, but it is worth noting in a government context: institutional knowledge is concentrated in very few hands.
2 of 15 committers made a single commit and never returned. 67% touched only one repository. A public organisation attracts drive-by contributors, and the numbers show GDS accepted them.
Weekday work dominates, as you would expect from a civil-service delivery organisation with a strong stated position on sustainable pace. The residual weekend activity is concentrated in incident response and personal-time open-source work.
0 commits come from 0 automated identities - overwhelmingly Dependabot. GOV.UK's documented policy is to auto-merge dependency upgrades where possible, and the commit history shows that policy working at scale.
In 0 of the 1 live repositories with meaningful history, one individual wrote more than half of all commits. That is the ordinary shape of a large estate with many small tools, but it is worth knowing which ones - they are listed below.
Who writes the code
Top 25 committers by commit count
Humans only. Bots are charted separately below.
Who they work for
Committers by email domain
A rough proxy for employment. GDS staff use gov.uk addresses; GitHub's privacy setting hides the rest, and a long tail commit from personal accounts.
Churn and tenure
Distinct people committing each year
Counting anyone whose active span covers that year. The shape tracks GDS headcount and the delivery peaks of the mid-2010s.
How long people stay
Span from first to last commit anywhere in the org.
When the work happens
Commits by hour of day
Author local time as recorded in the commit. A clean single-peak working day.
Commits by day of week
Reading the pattern
Weekend commits account for 5.1% of all activity. The working day peaks between 10:00 and 16:00 with a visible lunch dip, and tails off sharply after 18:00 - the signature of an organisation that does not run on overtime.
Concentration of knowledge
For each live repository with more than fifty human commits, how many people it takes to account for over half of them. One is a genuine key-person risk; four or more is a healthily shared codebase. 1 repositories qualify for the measure.
Bus factor across live repositories
Fewest contributors accounting for more than half the commits.
Where a single person wrote most of it
Share of commits by the largest single contributor, for the most concentrated repositories.
The machines
Top automated committers
Dependabot dominates. GOV.UK runs an in-house govuk-dependabot-merger service that auto-merges qualifying upgrade PRs, which is why bot commit volume is so high.
Top 60 committers
| Committer | Email type | Commits | Repos | First | Last | Years |
|---|---|---|---|---|---|---|
| sichunlam | github-private | 432 | 3 | 2017-05 | 2025-07 | 8.2 |
| Lam | gov.uk | 215 | 2 | 2017-05 | 2020-03 | 2.9 |
| Coventry City Council | github-private | 83 | 1 | 2022-12 | 2024-02 | 1.2 |
| Viviana Sagredo | github-private | 69 | 1 | 2022-06 | 2023-04 | 0.9 |
| lekshmiremadevi | github-private | 67 | 1 | 2020-04 | 2021-01 | 0.8 |
| cvbmc220 | github-private | 59 | 1 | 2021-02 | 2022-05 | 1.2 |
| holgatetom | github-private | 57 | 2 | 2020-08 | 2025-12 | 5.4 |
| HRowthorn | github-private | 46 | 1 | 2020-11 | 2021-08 | 0.8 |
| sichunlam | personal | 32 | 3 | 2019-07 | 2021-10 | 2.3 |
| Rukzana | github-private | 22 | 1 | 2024-01 | 2024-05 | 0.3 |
| Mullardman | github-private | 11 | 1 | 2020-11 | 2020-12 | 0.0 |
| lukaszmilton10 | github-private | 7 | 1 | 2020-06 | 2020-07 | 0.1 |
| JatinderCCC | github-private | 6 | 2 | 2024-03 | 2025-08 | 1.5 |
| justanewone | github-private | 1 | 1 | 2022-12 | 2022-12 | 0.0 |
| shivaangini | github-private | 1 | 1 | 2022-03 | 2022-03 | 0.0 |