Rising.
Repos under 5,000 stars that are gaining stars fastest. The same ranking as the main tables, in a table of its own.
Today: under 5,000 stars
This week: under 5,000 stars
How a repo gets here
Every day we read two public GitHub searches, both limited to repos with 1,000 to 4,999 stars, no forks and no archived repos: repos created in the last year, and repos pushed since yesterday. We save each repo's star count and subtract two saved counts. That is all the ranking does.
gained = stars_latest - stars_24_hours_earlier
# net of unstars; a repo that lost stars is not listed
The main tables work the same way and hold repos from 5,000 stars; How we rank has the full method and a worked example. A repo that reaches 5,000 stars moves there.
What Rising can't see
- Repos under 1,000 stars. About 100,000 repos have 500 to 4,999 stars, too many to read every day.
- Repos created more than a year ago that were not pushed since yesterday. A quiet old repo that suddenly takes off is not found.
- A repo's first day. It has no count from yesterday, so its first gain appears the day after.
- Bought stars. We can't tell a real surge from a bought one, here or in the main tables.
Questions
What is the Rising table?
It lists GitHub repos with 1,000 to 4,999 stars, ranked by the stars they gained: their count in the latest daily snapshot minus their count 24 hours before (and 7 days before, once we have a week of data). It is the same subtraction as the main rankings, on repos that are too small for them.
Why is it a separate table?
The main rankings track every repo from 5,000 stars. A repo under that line is found by a different search and watched in a different store, so we keep it in its own table with its own label instead of mixing two kinds of repo in one list. The main tables do not change when Rising changes.
Why is a repo I know missing?
Rising only sees repos with at least 1,000 stars that were created in the last year or pushed since yesterday. A repo with fewer stars, or an old repo with no push yesterday, is not in the set. A repo we see for the first time today has no count from yesterday, so its first gain shows up tomorrow.
Where do the numbers come from?
GitHub's public repository search, read once a day, read-only. We save each repo's star count and subtract two saved counts. You can redo any row by opening the repo on GitHub on two days and subtracting.
What happens when a Rising repo reaches 5,000 stars?
It moves to the main rankings, where it is ranked with every other repo from 5,000 stars, and leaves this table.