From Fedora Project Wiki

(Link with modern usage stats report.)
 
(133 intermediate revisions by 16 users not shown)
Line 1: Line 1:
{{admon/important|''See usage statistical report on [[Hardware probe|this page]].''}}
{{needs love}}
{{Admon/warning | This page is no longer being updated. | The methodologies here don't really work after several infrastructure changes over the last couple of years. This page needs to be revamped significantly to be meaningful. --[[User:Pfrields|pfrields]] ([[User talk:Pfrields|talk]]) 12:30, 20 August 2015 (UTC)}}
{{autolang|base=yes}}
{{autolang|base=yes}}


Line 7: Line 10:
Lots of different people.  Millions, in fact -- many millions!  You can see some notable examples [[Overview#Our_Community|on the Overview page]].
Lots of different people.  Millions, in fact -- many millions!  You can see some notable examples [[Overview#Our_Community|on the Overview page]].


== Fedora 15 (Lovelock) ==
=== Update Server Data ===


Lovelock ''was released'' on 2011-05-24. (Pithiness coming soon.)
==== Total repository connections ====
 
=== Yum Data ===
 
==== Current release connections ====
{|
|-
!colspan="5" | Connections to yum
|-
| Week || Dates || New Unique IPs || Total Unique IPs || Total compared to [[Legacy_statistics#Fedora_14_.28Laughlin.29|Laughlin]]
|-
| Week 1-4+ || 2011-05-24 -- 2011-06-23 ||    452,948 ||    452,948 ||  -- <ref name="interrupted">The normal weekly count was unintentionally interrupted by a log host rebuild. However, the cumulative count is still correct and can be maintained normally for the remainder of the release cycle.</ref>
|-
| Week 5-6  || 2011-06-24 -- 2011-07-04 ||    156,244 ||    609,192 ||  -- <ref name="interrupted"/>
|-
| Week  7  || 2011-07-05 -- 2011-07-11 ||    69,303 ||    678,495 || 107%
|-
| Week  8  || 2011-07-12 -- 2011-07-18 ||    82,583 ||    761,078 || 105%
|-
| Week  9  || 2011-07-19 -- 2011-07-23 ||    84,574 ||    845,652 || 105%
|-
| Week  10  || 2011-07-24 -- 2011-07-30 ||    83,585 ||    929,237 || 104%
|}
 
<!-- this link is out of date. --><!-- [[File:Yum-data-statistics.png|700px]] -->
 
<references />


==== Total repository connections ====
{|
!colspan="9" | Total connections to repositories
|-
| Fedora 7 || Fedora 8 || Fedora 9 || Fedora 10 || Fedora 11 || Fedora 12 || Fedora 13 || Fedora 14 || Fedora 15
|-
| 4,318,922 || 7,031,156 || 4,001,488 || 4,637,195 || 4,793,857 || 5,114,456 || 4,032,714 || 3,684,805 || 845,652
|-
| colspan="9" align=center | ''TOTAL UNIQUE IPs: 32,669,374''<br/>This total is not a simple sum of the counts above. It is a list of unique IP addresses that have connected across all releases.
|}




Line 52: Line 19:
{{admon/note|Currently, there is no reliable way to determine the total number of Linux users, or even count the total number of users of any Linux distribution which does not have a mandatory per user registration process.|Anyone who tells you otherwise may be misinformed, dishonest, or trying to sell you something.
{{admon/note|Currently, there is no reliable way to determine the total number of Linux users, or even count the total number of users of any Linux distribution which does not have a mandatory per user registration process.|Anyone who tells you otherwise may be misinformed, dishonest, or trying to sell you something.


The numbers above do not represent a total user count for Fedora.  We are counting IP addresses because that is the only thing we can reliably measure.  ''OEM shipping numbers, application popularity contests, and other "estimates" all have intrinsic error factors as well.''  Without a 1:1 sales figure or registration process, user counting is a difficult problem to solve.  Our [[#Accuracy_of_metrics|accuracy]] section below has a discussion of what's wrong with IP number counts, too.
The numbers above do not represent a total user or usage count or overall download count for Fedora.  Correlation of the numbers between different releases of Fedora is difficult due to the way our infrastructure is set up, and in part because the method we have used count IP connections has changed over time. We are counting IP addresses because they are the only statistic we can reliably measure.  ''Similarly, OEM shipping numbers, application popularity contests, and other "estimates" all have intrinsic error factors.''  Without a 1:1 sales figure or registration process, user counting is a difficult problem to solve.  Our [[#Accuracy_of_metrics|accuracy]] section has more information on this issue.  


When you hear someone estimate a count of users, the right questions to ask are:
When you hear of a user estimate, you should ask the following questions:
; Precisely how did you arrive at this number?
; How did you arrive at this number?
: For Fedora the answers are located at [[Statistics/Commands|this page]].
: For Fedora, the answers are located [[Statistics/Commands|here]].
; Where is the data located, so someone can duplicate the process?
; Where is the data located, to allow process verification?
: Although you do need appropriate access to our logging hosts to gather the data, our [[Infrastructure]] team does welcome people who want to participate in their work.}}
: Accessing Fedora's data requires appropriate access rights to our logging hosts, but as a community project, our [[Infrastructure]] team welcomes new contributors, allowing others to duplicate our analysis. }}


=== Accuracy of metrics ===
=== Accuracy of metrics ===
Line 67: Line 34:
# Users who are behind NAT or corporate proxies will not be counted at all.
# Users who are behind NAT or corporate proxies will not be counted at all.


The anecdotal evidence that we receive from different groups, companies, and organizations suggests that group (2) is significantly larger than group (1).  As such, we believe that the true numbers in the field are higher than the numbers on this page.
The anecdotal evidence that we receive from different groups, companies, and organizations suggests that group (2) is significantly larger than group (1).  As such, we believe that the true numbers in the field are higher than the numbers on this page.   
 
For a more thorough discussion and enlightening information about the measurement of Fedora's userbase, refer to [http://jspaleta.livejournal.com/42464.html this blog entry] by [[User:Jspaleta | Jef Spaleta]].
 
=== Why is the Fedora 8 count so high? ===
 
We believe this is due to the high concentration of Fedora 8 images on Amazon EC2We may see additional changes with Fedora 14 and later as that release is expected to be [[Fedora_14_on_EC2#Interview_with_Justin_Forbes|available on EC2]] on release day.
 
== Smolt Data ==
 
[http://smolt.fedoraproject.org Smolt]  is Fedora's hardware profiling project.  It is an '''opt-in''' database that tracks unique installations of Fedora, and various details about that install.  People who install to runlevel 5 will be able to opt-in during firstboot, but people who install to runlevel 3 will have to install the smolt client via yum and manually register.
 
== BitTorrent ==
 
The following table shows the number of downloads that have been made over BitTorrent. This table shows downloads only through trackers connected to the [http://torrent.fedoraproject.org official torrent server].
 
{|
|-
!colspan="6" | Downloads from bittorrent (as of 2011-04-12)
|-
| Sulphur (F9) || Cambridge (F10) || Leonidas (F11)<ref>The statistics for BitTorrent were interrupted on or about 2009-09-04 Fedora 11 because of a torrent server reset. Totals at that time are added to the current totals (F10: 306,696, F11: 169,832).</ref> || Constantine (F12) || Goddard (F13) || Laughlin (F14)
|-
| 443,932 || 523,776 || 364,808 || 315,349 || 158,570 ||  157,251
|-
| colspan=6 | ''NOTE: these numbers seem incomplete and need to be checked against torrent records. --[[User:Pfrields|pfrields]] 11:35, 13 April 2011 (UTC)''
|}
 
<references/>
 
== Direct Download ==
 
The following table shows the number of direct downloads of {{FedoraVersion|long}} media '''from unique IP addresses''', as shown in the web proxy logs.  The actual number of raw downloads tends to be much higher.
 
{|
|-
!colspan="5" | Direct downloads from the web<br/>
(includes a count of DVD, Live, CD disc 1, and netinst ISOs downloaded directly from the web)
|-
! Week !! Dates !! Downloads this week !! Total downloads !! Total compared to [[Legacy_statistics#Fedora_14_.28Laughlin.29|Laughlin]]
|-
| Week  1  || 2011-05-24 -- 2011-05-30 ||  224,960 ||  224,960 || 148%
|-
| Week  2  || 2011-05-31 -- 2011-06-06 ||  105,883 ||  330,843 || 133%
|-
| Week  3  || 2011-06-07 -- 2011-06-13 ||    86,137 ||  416,980 || 128%
|-
| Week  4  || 2011-06-14 -- 2011-06-20 ||    85,315 ||  502,295 || 127%
|-
| Week  5  || 2011-06-21 -- 2011-06-27 ||    66,305 ||  568,600 || 125%
|-
| Week  6  || 2011-06-28 -- 2011-07-04 ||    72,389 ||  640,989 || 125%
|-
| Week  7  || 2011-07-05 -- 2011-07-11 ||    68,984 ||  709,973 || 112%
|-
| Week  8  || 2011-07-12 -- 2011-07-18 ||    67,079 ||  777,052 || 126%
|-
| Week  9  || 2011-07-19 -- 2011-07-25 ||    69,009 ||  846,061 || 127%
|}
 
== fedoraproject.org unique visitors ==
 
{|
|-
!colspan="2" | fedoraproject.org visitors
|-
| Period || Unique IPs
|-
| January 2008  || 785,583
|-
| February 2008  || 782,043
|-
| March 2008    || 830,078
|-
| April 2008    || 723,776
|-
| May 2008      || 1,128,115
|-
| June 2008      || 825,912
|-
| July 2008      || 751,521
|-
| August 2008    || 773,209
|-
| September 2008 || 1,461,490
|-
| October 2008  || 1,351,477
|-
| November 2008  || 1,993,547
|-
| December 2008  || 1,873,217
|-
| January 2009  || 1,815,135
|-
| February 2009  || 1,697,108
|-
| March 2009    || 1,816,077
|-
| April 2009    || 1,749,571
|-
| May 2009      || 1,996,040
|-
| June 2009      || 2,305,280
|-
| July 2009      || 1,929,651
|-
| August 2009    || 1,964,544
|-
| September 2009 || 1,859,708
|-
| October 2009  || 1,996,649
|-
| November 2009  || 2,423,626 (all-time monthly high)
|-
| December 2009  || 1,969,482
|-
| January 2010  || 1,829,467
|-
| February 2010  || 1,628,215
|-
| March 2010    || 1,728,081
|-
| April 2010    || 1,585,905
|-
| May 2010      || 2,197,213
|-
| June 2010      || 1,853,870
|-
| July 2010      || 1,808,418
|-
| August 2010    || 1,589,666
|-
| September 2010 || 1,642,009
|-
| October 2010  || 1,903,774
|-
| November 2010  || 2,086,089
|-
| December 2010  || 1,645,695
|-
| January 2011  || 1,633,474
|-
| February 2011  || 1,493,936
|-
| March 2011    || 1,636,965
|-
| April 2011    || 1,776,393
|-
| May 2011      || 2,251,895
|-
| June 2011      || 2,097,498
|-
| July 2011      || 1,814,156
|}
 
[[File:Fedoraproject-org-statistics.png]]
 
== Edits to Fedora wiki ==
 
{|
|-
!colspan="3" | Fedora wiki edits
|-
| Period || Total edits || Edits from unique IPs
|-
| October 2007  || 921  || 426
|-
| November 2007  || 1,090 || 472
|-
| December 2007  || 807  || 378
|-
| January 2008  || 1,170 || 580
|-
| February 2008  || 1,244 || 645
|-
| March 2008    || 1,280 || 705
|-
| April 2008    || 1,283 || 755
|-
| May 2008      || 6,043 || 766
|-
| June 2008      || 6,561 || 731
|-
| July 2008      || 4,112 || 528
|-
| August 2008    || 2,226 || 423
|-
| September 2008 || 4,738 || 791
|-
| October 2008  || 3,848 || 617
|-
| November 2008  || 4,619 || 715
|-
| December 2008  || 4,627 || 751
|-
| January 2009  || 6,327 || 739
|-
| February 2009  || 7,444 || 764
|-
| March 2009    || 8,650 || 973
|-
| April 2009    || 5,294 || 844
|-
| May 2009      || 5,191 || 928
|-
| June 2009      || 5,337 || 858
|-
| July 2009      || 5,373 || 807
|-
| August 2009    || 4,864 || 702
|-
| September 2009 || 5,796 || 868
|-
| October 2009  || 5,407 || 782
|-
| November 2009  || 5,665 || 957
|-
| December 2009  || 4,058 || 653
|-
| January 2010  || 5,319 || 711
|-
| February 2010  || 12,384 || 761
|-
| March 2010    || 12,176 || 834
|-
| April 2010    || 13,874 || 1,001
|-
| May 2010      || 17,809 || 871
|-
| June 2010      || 13,879 || 672
|-
| July 2010      || 14,274 || 793
|-
| August 2010    || 13,657 || 791
|-
| September 2010 || 12,144 || 788
|-
| October 2010  || 12,400 || 888
|-
| November 2010  || 12,447 || 887
|-
| December 2010  ||  7,458 || 627
|-
| January 2011  ||  9,768 || 771
|-
| February 2011  || 11,920 || 950
|-
| March 2011    || 11,909 || 882
|-
| April 2011    || 10,774 || 835
|-
| May 2011      ||  5,628 || 917
|-
| June 2011      ||  6,129 || 767
|-
| July 2011      ||  5,592 || 725
|}
 
== Contributors ==
<!-- This chart is pretty ugly.
{|
! Month !! CLA
|-
| 2005-05 || 122
|-
| 2005-06 || 149
|-
| 2005-07 || 174
|-
| 2005-08 || 206
|-
| 2005-09 || 225
|-
| 2005-10 || 238
|-
| 2005-11 || 250
|-
| 2005-12 || 268
|-
| 2006-01 || 289
|-
| 2006-02 || 319
|-
| 2006-03 || 372
|-
| 2006-04 || 433
|-
| 2006-05 || 486
|-
| 2006-06 || 560
|-
| 2006-07 || 610
|-
| 2006-08 || 654
|-
| 2006-09 || 693
|-
| 2006-10 || 739
|-
| 2006-11 || 794
|-
| 2006-12 || 844
|-
| 2007-01 || 890
|-
| 2007-02 || 949
|-
| 2007-03 || 1009
|-
| 2007-04 || 1093
|-
| 2007-05 || 1182
|-
| 2007-06 || 1298
|-
| 2007-07 || 1403
|-
| 2007-08 || 1511
|-
| 2007-09 || 1610
|-
| 2007-10 || 1693
|-
| 2007-11 || 1794
|-
| 2007-12 || 1918
|-
| 2008-01 || 2000
|-
| 2008-02 || 2119
|-
| 2008-03 || 2222
|-
| 2008-04 || 2393
|-
| 2008-05 || 2713
|-
| 2008-06 || 3641
|-
| 2008-07 || 4431
|-
| 2008-08 || 5173
|-
| 2008-09 || 5869
|-
| 2008-10 || 6724
|-
| 2008-11 || 7345
|-
| 2008-12 || 8127
|-
| 2009-01 || 8839
|-
| 2009-02 || 9565
|-
| 2009-03 || 10229
|-
| 2009-04 || 10957
|-
| 2009-05 || 11558
|-
| 2009-06 || 12323
|-
| 2009-07 || 13254
|-
| 2009-08 || 13924
|-
| 2009-09 || 14604
|-
| 2009-10 || 15241
|}
-->
[[Image:Accounts_2010-02.jpg]]
 
In April 2008, the [[Account_System| FAS (Fedora Account System)]] process was converted to a click-through process instead of a manual process of GPG signing and fax/email.
 
=== Translations ===
 
Details about the progress of the [[L10N|localisation]] can be found in our [https://translate.fedoraproject.org/ transifex] instance.
 
=== Ambassadors ===
 
Information about the [https://fedorahosted.org/fama/wiki/AmbassadorMetrics number of Fedora Ambassadors].
 
=== Translators ===
 
Information on the [http://dimitris.glezos.com/weblog/2009/03/08/flp-advancement-study growth and numbers of Fedora Translators].
 
== Maps ==
<!-- A [http://fedoraproject.org/maps/wiki-edits/ map]  of wiki edits worldwide. -->
<!-- Maps not working again yet.  Huzaifa is working on this. -->


A variety of [[maps | maps]] showing various statistics.
For a more thorough discussion and enlightening information about the measurement of Fedora's userbase, refer to [http://jspaleta.livejournal.com/42464.html this blog entry] by [[User:Jspaleta | Jef Spaleta]]


== Charts ==
== Charts ==

Latest revision as of 07:12, 11 March 2020

See usage statistical report on this page.
This page needs some love
This page should be revised or reconstructed to be more helpful. Problems may include being out of step with current team or project status or process.
This page is no longer being updated.
The methodologies here don't really work after several infrastructure changes over the last couple of years. This page needs to be revamped significantly to be meaningful. --pfrields (talk) 12:30, 20 August 2015 (UTC)

This page contains a variety of statistics about Fedora's usage. The Fedora Project believes that metrics are an important tool.

Who uses Fedora?

Lots of different people. Millions, in fact -- many millions! You can see some notable examples on the Overview page.

Update Server Data

Total repository connections

Currently, there is no reliable way to determine the total number of Linux users, or even count the total number of users of any Linux distribution which does not have a mandatory per user registration process.
Anyone who tells you otherwise may be misinformed, dishonest, or trying to sell you something.

The numbers above do not represent a total user or usage count or overall download count for Fedora. Correlation of the numbers between different releases of Fedora is difficult due to the way our infrastructure is set up, and in part because the method we have used count IP connections has changed over time. We are counting IP addresses because they are the only statistic we can reliably measure. Similarly, OEM shipping numbers, application popularity contests, and other "estimates" all have intrinsic error factors. Without a 1:1 sales figure or registration process, user counting is a difficult problem to solve. Our accuracy section has more information on this issue.

When you hear of a user estimate, you should ask the following questions:

How did you arrive at this number?
For Fedora, the answers are located here.
Where is the data located, to allow process verification?
Accessing Fedora's data requires appropriate access rights to our logging hosts, but as a community project, our Infrastructure team welcomes new contributors, allowing others to duplicate our analysis.

Accuracy of metrics

The numbers above, for yum, represent unique IP addresses that reach our update server, not simply downloads. We believe it is reasonable to equate a "new IP address checking in" with "a new installation of Fedora", with the following caveats:

  1. Users who have dynamic IP addresses will likely be counted multiple times, which inflates the number by some amount.
  2. Users who are behind NAT or corporate proxies will not be counted at all.

The anecdotal evidence that we receive from different groups, companies, and organizations suggests that group (2) is significantly larger than group (1). As such, we believe that the true numbers in the field are higher than the numbers on this page.

For a more thorough discussion and enlightening information about the measurement of Fedora's userbase, refer to this blog entry by Jef Spaleta

Charts

Some charts are available reflecting some of the information shown on this page. Up-to-date charts are available at the Fedora community portal.

Legacy Statistics

Older releases and years are on the Legacy statistics page.

Commands

Information on how these statistics are gathered is found here.