File talk:Usage share of web browsers (Source StatCounter).svg

Why is the legend in front of one of the stats lines? I think it would be better to move it but I don't know how. Good point for the next graph! --Rderijcke (talk) 10:21, 2 August 2011 (UTC)

Noticed that too, I can't tell how fast firefox rose where the legend is71.15.125.143 (talk) 03:58, 23 January 2012 (UTC)

Yes, you definitely should move the legend away from the lines! 83.78.125.12 (talk) 15:32, 5 February 2012 (UTC)

 Done Litehacker (talk) 19:32, 1 April 2012 (UTC)

Inflation exceeds 12% as of September 2013.

df <- data.frame (explorer=browser.ie, firefox=browser.firefox, chrome=browser.chrome, safari=browser.safari, opera=browser.opera, mobile=browser.mobile)
require (plyr)
ddply (df, 1:5, sum)

   explorer firefox chrome safari opera     V1
1     27.72   19.76  41.38   7.96  1.00 112.44
2     29.30   20.87  38.07   8.50  1.17 112.35
3     29.71   20.06  39.15   8.00  1.01 111.83
4     29.82   21.34  37.09   8.60  1.22 112.42
5     30.71   21.42  36.52   8.29  1.19 112.26
6     30.78   21.89  36.42   7.92  1.26 112.82
7     31.23   22.37  35.72   7.83  1.39 111.62
8     32.04   23.73  33.81   7.12  1.72 109.51
9     32.08   22.32  34.77   7.81  1.63 110.91
10    32.12   25.55  32.43   7.09  1.77 109.07
11    32.31   24.56  32.76   7.00  1.77 108.80
12    32.70   22.40  34.21   7.70  1.61 110.65
13    32.85   22.85  33.59   7.39  1.63 110.09
14    34.07   24.87  31.23   7.13  1.72 108.60
15    34.81   24.98  30.87   6.72  1.78 108.15
16    35.75   24.88  29.84   6.77  2.02 107.79
17    37.45   24.78  28.40   6.62  1.95 107.69
18    38.65   25.27  27.27   6.08  1.98 107.29
19    40.18   26.39  25.00   5.93  1.81 105.86
20    40.63   25.23  25.69   5.92  1.82 106.24
21    41.66   26.79  23.61   5.60  1.72 106.12
22    41.89   27.49  23.16   5.19  1.67 106.52
23    42.45   27.95  22.14   5.16  1.66 106.38
24    43.58   28.34  20.65   5.07  1.74 105.91
25    43.87   29.29  19.36   5.01  1.84 105.12
26    44.52   29.67  18.29   5.04  1.91 104.64
27    45.11   29.98  17.37   5.02  1.97 104.15
28    45.44   30.37  16.54   5.08  2.00 103.88
29    46.00   30.68  15.68   5.09  2.00 103.75
30    46.94   30.76  14.85   4.79  2.07 103.51
31    48.16   31.17  13.35   4.70  2.01 103.41
32    49.21   31.24  12.39   4.56  2.00 103.21
33    49.87   31.50  11.54   4.42  2.03 102.86
34    51.34   31.09  10.76   4.23  1.88 102.51
35    52.68   30.69   9.88   4.09  1.91 102.11
36    52.77   31.64   8.61   4.14  1.96 101.44
37    52.86   31.15   9.24   4.07  1.91 101.80
38    53.26   31.74   8.06   4.23  1.82 101.29
39    54.44   31.27   7.29   4.16  1.97 101.09
40    54.50   31.83   6.71   4.08  1.97 100.81
41    55.25   31.64   6.04   3.76  2.00 100.25
42    55.72   31.97   5.45   3.48  2.06  99.96
43    56.57   32.21   4.66   3.67  2.02 100.34
44    57.96   31.82   4.17   3.47  1.88 100.45
45    58.37   31.34   3.69   3.28  2.62 100.42
46    58.69   31.28   3.38   3.25  2.67 100.39
47    59.49   30.26   2.80   2.91  3.46  99.86
48    60.11   30.50   3.01   3.02  2.64 100.33
49    61.88   29.67   2.07   2.75  2.96 100.19
50    62.09   28.75   2.42   2.65  3.23 100.00
51    62.52   29.40   1.73   2.73  2.94 100.12
52    64.43   27.85   1.52   2.59  2.95 100.03
53    65.41   27.03   1.38   2.57  2.92  99.98
54    67.16   25.77   1.03   3.00  2.86     NA
55    67.68   25.54   1.02   2.91  2.69     NA
56    67.84   25.23   1.21   2.41  2.83 100.12
57    68.14   25.27   0.93   2.49  3.01     NA
58    68.57   26.14     NA   3.30  1.78     NA
59    68.91   26.08     NA   2.99  1.83     NA

I was wondering about peak choice.

 H <- function (v) { p<-v/sum(v); -sum(p*log2(p)) }

ddply (df, 1:6, H)$V1
  2.183331 2.190255 2.216613 2.224303 2.218107 2.219387 2.215093 2.218628
  2.214239 2.210555 2.205386 2.201392 2.201056 2.190397 2.175899 2.177155
 2.165872 2.145324 2.108127 2.103482 2.086247 2.079735 2.071019 2.051073
 2.030112 2.011639 1.991809 1.979913 1.965589 1.942024 1.911546 1.883276
 1.856113 1.815062 1.777646 1.755233 1.739772 1.717507 1.688478 1.663445
 1.626796 1.589395 1.562894 1.522660 1.525238 1.513397 1.475864 1.491793
 1.431639 1.416333 1.391337 1.350074 1.328980 1.277658       NA       NA
       NA       NA       NA

As I suspected, we recently crested peak diversity as estimated by Shannon entropy.

Postscript The NA problem can be handled like this:

H <- function (v) { p<-v/sum(v, na.rm=TRUE); -sum(p*log2(p), na.rm=TRUE) }

— MaxEnt 10:35, 15 September 2013 (UTC)

What is this line? Based on previous versions calling it "mobile", it seems to be the sum of all mobile users. Why not split them by browser, as the name of the graph suggests? --mfb (talk) 12:58, 25 May 2017 (UTC)