Closed andymeneely closed 10 years ago
Here's what a histogram from a few weeks ago looked like. Haven't gotten to the scatter plot yet.
@@@ Max Churn per CodeReview Histogram @@@
value |---------------------------------------------------------------------------------------------------------| count
0 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@| 49460
10 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ | 18459
20 |@@@@@@@@@@@@@@@@@@@@@@ | 10727
30 |@@@@@@@@@@@@@@@@ | 7666
40 |@@@@@@@@@@@@ | 5956
50 |@@@@@@@@@@ | 4838
60 |@@@@@@@@ | 3934
70 |@@@@@@@ | 3577
80 |@@@@@@ | 2993
90 |@@@@@ | 2749
100 |@@@@@ | 2452
110 |@@@@ | 2243
120 |@@@@ | 1937
130 |@@@ | 1698
140 |@@@ | 1673
150 |@@@ | 1489
160 |@@ | 1413
170 |@@ | 1257
180 |@@ | 1266
190 |@@ | 1096
200 |@@ | 1037
210 |@@ | 999
220 |@@ | 978
230 |@ | 922
240 |@ | 848
250 |@ | 788
260 |@ | 790
270 |@ | 681
280 |@ | 646
290 |@ | 636
300 |@ | 584
310 |@ | 590
320 |@ | 524
330 |@ | 520
340 |@ | 528
350 |@ | 515
360 | | 466
370 | | 434
380 | | 404
390 | | 440
400 | | 367
410 | | 391
420 | | 398
430 | | 345
440 | | 347
450 | | 327
460 | | 308
470 | | 330
480 | | 308
490 | | 335
500 | | 322
510 | | 248
520 | | 277
530 | | 250
540 | | 257
550 | | 239
560 | | 264
570 | | 241
580 | | 228
590 | | 227
600 | | 217
610 | | 205
620 | | 215
630 | | 194
640 | | 185
650 | | 171
660 | | 183
670 | | 152
680 | | 176
690 | | 157
700 | | 149
710 | | 136
720 | | 155
730 | | 130
740 | | 153
750 | | 151
760 | | 138
770 | | 118
780 | | 150
790 | | 113
800 | | 152
810 | | 110
820 | | 117
830 | | 138
840 | | 122
850 | | 110
860 | | 100
870 | | 102
880 | | 87
890 | | 97
900 | | 116
910 | | 93
920 | | 108
930 | | 105
940 | | 92
950 | | 63
960 | | 88
970 | | 84
980 | | 85
990 | | 84
Total |---------------------------------------------------------------------------------------------------------| 151523
@kayladavis Let's follow up with this question. Use something like Excel or R to construct a 2D plot of max_churn
and comments
as well as max_churn
and messages
over CodeReview
s.
Don't worry about automating plots just yet (although that might be on the horizon)
Looks like there was nothing obvious for this threshold, so this is done.
Look at the histogram of Max Churn and graph it against number of comments and messages. Does the participation drop off when churn gets higher?
Is there a clear threshold for when developers "burn out"? If so, determine that threshold and make a metric out of it. Maybe
CodeReview#burnout?