Github stats

Publishing code on Github or other open hosting is often referred to as such life-giving euthanasia, after which patches, fixes, problem reports, and other third-party commits will flow. I still have the opposite experience. In a couple of my relatively visible projects in general, no one and nothing really brought, not to mention the other projects. Moreover, several times I made very serious changes / improvements to other people's projects, but their authors froze and did not accept the changes. Based on this, I developed a theory: they say, all this openness and collaboration are beautiful words, really everyone is sawing something of their own and earning money or PR to someone else is completely unwilling. To test the theory, I counted some statistics on all repositories on Github, which have more than 700 stars, a little more than 4 thousand.

Legend

The language is the main language of the project, None means that the project consists only of text (.txt, .md, without the extension) or binary files.
Commits - the total number of commits made by hundreds of the most active authorized people.
Additions - the total difference of added and deleted lines for all commits made by the hundred most active authorized people.
Changes - the total amount of added and deleted lines for all commits made by hundreds of the most active authorized people.
Authors - the number of people who in the amount made more than 90% of commits, and / or added more than 90% of the lines, and / or made more than 90% of the changes. At least two of the three conditions must be met.
')
The main drawback of the study is the so-called. unauthorized commits are not counted at all. And there are quite a few of them; in many projects, almost all commits are unauthorized. An extreme case - there are more than 150 thousand commits in the WebKit / webkit project, but there are no authorized ones :( So many of the results below are very inaccurate / incomplete.

Below you will see many projects with a negative number of additions. This means that the main development goes unauthorized commits, and authorized users only delete the code :)

The second drawback is that commits from authorized people who are not in the first hundred by the number of commits are not taken into account. However, this affects the units of projects, and that is not much, because in the “tail” are usually people who have made one single-line commit. Probably, the record holder in the number of authorized contributors is the Homebrew / homebrew project, 4,146 people.

And now a bunch of tables:

The authors	Number of projects
0	10 - projects without authorized commits, for example. nginx / nginx , v8 / v8 , or too new, for example. ericjohnson / canabalt-ios
one	1534 - 36.7%
2	766 - 18.3%
3	449 - 10.7%
four	264 - 6.3%
five	181 - 4.3%
6	151 - 3.6%
7	105 - 2.5%
eight	83 - 2.0%
9	63 - 1.5%
ten	75 - 1.8%
eleven	54 - 1.3%
12	42 - 1.0%
13	50 - 1.2%
14	44 - 1.1%
15	24 - 0.6%
sixteen	31 - 0.7%
17	21 - 0.5%
18	22 - 0.5%
nineteen	16 - 0.4%
20	14 - 0.3%
21	10 - 0.2%
22	19 - 0.5%
23	17 - 0.4%
24	9 - 0.2%
25	12 - 0.3%
26	8 - 0,2%
27	11 - 0.3%
28	8 - 0,2%
29	3 - 0.1%
thirty	8 - 0,2%
31	6 - 0.1%
32	4 - 0.1%
33	7 - 0.2%
34	4 - 0.1%
35	7 - 0.2%
36	6 - 0.1%
37	2 - 0.0%
38	3 - 0.1%
39	1 - 0.0%
40	2 - 0.0%
41	2 - 0.0%
42	2 - 0.0%
44	4 - 0.1%
45	1 - 0.0%
46	1 - 0.0%
48	4 - 0.1%
50	2 - 0.0%
51	2 - 0.0%
52	3 - 0.1%
55	1 - 0.0%
57	1 - 0.0%
58	2 - 0.0%
59	1 - 0.0%
62	1 - 0.0%
63	2 - 0.0%
64	1 - 0.0%
65	1 - 0.0%
66	1 - 0.0%
68	1 - 0.0%
Total	4179

In general, my thesis is more likely not justified, only 36% of popular projects are developed alone. Although here how to look, because these are only projects with 700 stars!

Top 20 projects by number of authors

Project	Tongue	Asterisks	Commits	Additions	Changes	The authors
github / gitignore	None	18204	372	790	3870	68
vhf / free-programming-books	None	28449	1101	1123	24035	67
raspberrypi / linux	C	1696	28833	324302	2623244	65
purplecabbage / phonegap-plugins	Java	1075	835	-105528	1589276	64
torvalds / linux	C	15394	43786	507,000	4359572	63
openstack / nova	Python	1329	11074	220561	2649687	63
github / github-services	Ruby	1474	1615	-66297	468497	62
github / hubot-scripts	CoffeeScript	2486	938	17244	32354	59
Shopify / active_merchant	Ruby	2784	1373	25326	161334	58
TechEmpower / FrameworkBenchmarks	Php	1347	3140	8601	12601129	58
h5bp / server-configs	Javascript	1456	365	-131	29431	57
adambard / learnxinyminutes-docs	None	1772	846	43167	73191	55
kahun / awesome-sysadmin	None	7151	125	178	296	52
munin-monitoring / contrib	Perl	722	875	44090	337494	52
svenfuchs / rails-i18n	Ruby	2132	747	10982	82176	52
emberjs / data	Javascript	1999	1164	-49876	329034	51
braziljs / js-the-right-way	CSS	3538	172	-8325	48869	51
zendframework / zf2	Php	4843	9592	-1352588	9532926	50
CocoaPods / Specs	Ruby	1904	7323	319548	1026292	48
rails / docrails	Ruby	948	30360	172063	3529109	48

Top 20 projects by the number of commits

Project	Tongue	Asterisks	Commits	Additions	Changes	The authors
mono / mono	C #	2426	62264	132299	18585797	27
freebsd / freebsd	C	701	57479	-40239109	93143111	35
phpmyadmin / phpmyadmin	Php	945	56388	1708781	33424887	four
JetBrains / intellij-community	Java	1345	48870	2437614	8986452	22
Ffmpeg / ffmpeg	C	1250	46411	665608	3186750	25
torvalds / linux	C	15394	43786	507,000	4359572	63
plt / racket	Racket	710	32077	2037784	18497412	7
odoo / odoo	Python	1113	31110	-1274226	7244122	thirty
rails / rails	Ruby	23104	30372	172648	3529808	48
rails / docrails	Ruby	948	30360	172063	3529109	48
mono / monodevelop	C #	1122	29941	-254891	19631973	eight
raspberrypi / linux	C	1696	28833	324302	2623244	65
php / php-src	C	4656	28492	1160281	4475461	21
ceph / ceph	C ++	862	27577	-72820	3031696	eleven
ruby / ruby	Ruby	5904	27415	906097	3703481	eleven
Homebrew / homebrew	Ruby	19145	25179	44673	388961	25
rust-lang / rust	Rusty	6258	23984	500926	6043730	37
saltstack / salt	Python	3965	22543	422978	1815546	27
postgres / postgres	C	1051	22393	1632543	10019479	6
neo4j / neo4j	Java	1050	22280	1074573	9574625	sixteen

This and the two following tables were badly damaged due to unauthorized commits. In torvalds / linux , only 43,000 authorized from the first hundred authors remained from 470,000 commits , and the WebKit / webkit already mentioned, with its 150,000 commits, did not get here at all.

Top 20 projects by number of additions

Project	Tongue	Asterisks	Commits	Additions	Changes	The authors
cdnjs / cdnjs	Javascript	2171	3487	61274218	66501052	five
usrbinnc / netcat-cpi-kernel-module	C	797	68	37786452	113360024	2
johnmyleswhite / ML_for_Hackers	R	1407	72	26858328	79840760	2
Team Sass / modular-scale	CSS	858	214	20594705	41423603	one
jsdelivr / jsdelivr	Javascript	979	2037	14843627	23807661	9
shiffman / The-Nature-of-Code-archive	Javascript	877	1027	11632840	21615730	2
ariya / phantomjs	C ++	11199	1080	8994617	15191223	2
webscalesql / webscalesql-5.6	C ++	2348	65	7230744	10258858	one
cloudera / Impala	C ++	1052	1931	6994463	61862823	3
divegeek / uscode	Python	712	65	6534521	21863397	one
xamarin / XobotOS	Java	1176	one	5956527	5956527	one
mozilla b2g / gaia	Javascript	1724	18637	5368630	103734808	14
bundestag / gesetze	None	1159	347	4857181	4957541	one
ValveSoftware / source-sdk-2013	C ++	1275	60	4530235	13493049	2
guardian / frontend	Javascript	2264	17822	4331243	44952471	7
AppScale / appscale	Python	1772	1918	3893760	11496912	2
IronLanguages ​​/ main	Python	735	1343	3792378	16179138	four
roboguice / roboguice	Java	2129	164	3499491	3514803	one
mozilla / BrowserQuest	Javascript	4443	148	3390557	3394023	one
runtimejs / runtime	C ++	775	147	3243017	3698303	one

The first honest project here is ariya / phantomjs .

Top 20 projects by number of changes

Project	Tongue	Asterisks	Commits	Additions	Changes	The authors
usrbinnc / netcat-cpi-kernel-module	C	797	68	37786452	113360024	2
mozilla b2g / gaia	Javascript	1724	18637	5368630	103734808	14
freebsd / freebsd	C	701	57479	-40239109	93143111	35
johnmyleswhite / ML_for_Hackers	R	1407	72	26858328	79840760	2
cdnjs / cdnjs	Javascript	2171	3487	61274218	66501052	five
cloudera / Impala	C ++	1052	1931	6994463	61862823	3
sass / libsass	C ++	1805	1217	-10007815	50622323	9
guardian / frontend	Javascript	2264	17822	4331243	44952471	7
OpenELEC / OpenELEC.tv	Shell	1117	14043	232685	43384935	2
Team Sass / modular-scale	CSS	858	214	20594705	41423603	one
phpmyadmin / phpmyadmin	Php	945	56388	1708781	33424887	four
moodle / moodle	Php	947	21043	1465212	29927296	14
manuels / texlive.js	TeX	731	31	2220800	27001198	one
phalcon / cphalcon	C	4035	1746	-131973	25292871	ten
jsdelivr / jsdelivr	Javascript	979	2037	14843627	23807661	9
divegeek / uscode	Python	712	65	6534521	21863397	one
shiffman / The-Nature-of-Code-archive	Javascript	877	1027	11632840	21615730	2
sass / node-sass	CSS	971	1058	26714	20446906	eight
mono / monodevelop	C #	1122	29941	-254891	19631973	eight
facebook / hhvm	C ++	9676	9491	1602075	19429851	24

Top 7 projects by the number of stars with one author

Project	Tongue	Asterisks	Commits	Additions	Changes	The authors
FortAwesome / Font-Awesome	CSS	26381	579	147220	598986	one
bartaz / impress.js	Javascript	21409	183	2262	7022	one
blueimp / jQuery-File-Upload	Javascript	17414	759	10172	52544	one
jakubroztocil / httpie	Python	11638	506	6888	34684	one
nnnick / Chart.js	Javascript	11167	69	8514	26340	one
Prinzhorn / skrollr	Javascript	10217	527	7391	36179	one
octocat / Spoon-Knife	CSS	9778	3	46	48	one

Top 7 projects by the number of commits with one author

Project	Tongue	Asterisks	Commits	Additions	Changes	The authors
joeyh / git-annex	Haskell	752	9351	1827373	2333911	one
tekezo / Karabiner	C ++	750	8021	312620	1151542	one
ornicar / lila	Scala	701	7402	99107	651415	one
kraih / mojo	Perl	1316	6233	29862	385940	one
Atmosphere / atmosphere	Java	1855	6224	-484605	2995439	one
getsentry / sentry	Python	5649	5431	277348	1129168	one
facebook / pfff	Ocaml	948	4955	242737	1357931	one

So, the main maniac is joeyh, who made at least 10 thousand commits (judging by the repository, probably 20, half unauthorized), but did not wait for normal community support. It is symbolic that the project on Haskell.

Next - statistics separately by language.

Tongue	-	-
Javascript	1334 — 31,9%	3.6
Ruby	495 — 11,8%	8.5
Objective c	409 — 9,8%	2.4
Python	348 — 8,3%	5.9
Java	240 — 5,7%	4.7
CSS	231 — 5,5%	3.0
Php	199 — 4,8%	6.9
C	155 — 3,7%	5.5
None	132 — 3,2%	6,6
C ++	106 — 2,5%	4.9
Go	91 — 2,2%	3,4
Shell	82 — 2,0%	6,6
CoffeeScript	70 — 1,7%	4.9
VimL	50 — 1,2%	5.1
C #	41 — 1,0%	7.0
Scala	39 — 0,9%	6.4
Clojure	29 — 0,7%	5.6
Swift	15 — 0,4%	1.8
Perl	14 — 0,3%	5.9
Emacs Lisp	12 — 0,3%	6.5
TeX	10 — 0,2%	2.4
Haskell	9 — 0,2%	3.9
Erlang	8 — 0,2%	9.9
ActionScript	6 — 0,1%	2.2
Assembly	6 — 0,1%	18,8
R	5 — 0,1%	2.8
Lua	4 — 0,1%	7.5
Rust	4 — 0,1%	16.5
Powershell	3 — 0,1%	4.7
XSLT	3 — 0,1%	3.7
OCaml	3 — 0,1%	2.7
Objective-C++	3 — 0,1%	2.7
Elixir	3 — 0,1%	6,3
Groovy	2 — 0,0%	3.5
Matlab	2 — 0,0%	2.0
TypeScript	2 — 0,0%	11.0
Dart	2 — 0,0%	8.0
LiveScript	1 — 0,0%	3.0
Racket	1 — 0,0%	7.0
Scheme	1 — 0,0%	2.0
Vala	1 — 0,0%	3.0
Common Lisp	1 — 0,0%	6.0
Visual Basic	1 — 0,0%	1.0
Objective-J	1 — 0,0%	10.0
Io	1 — 0,0%	1.0
Mirah	1 — 0,0%	3.0
Julia	1 — 0,0%	21.0
IDL	1 — 0,0%	2.0
Nimrod	1 — 0,0%	8.0

Objective-C, CSS, Swift, TeX ActionScript , , , , - .

Ruby-. , , Ruby.

---

, : gist.github.com/leventov/9e00036ec2f3abbf4a55

, : docs.google.com/spreadsheets/d/16YRLzNzIfBAYXQUiPteDkCYgLL3eVM8qMSmiRQySFl0/edit?usp=sharing

JSON (18 ): vk.com/doc14882053_324898086 ( , '700.json' 'github')