Time Series on monthly statistics of forum (new users, new topics, new posts)

[tranthidung]

TIME SERIES ON MONTHLY STATISTICS OF FORUM

(NEW USERS, NEW TOPICS, NEW POSTS)

Hello all,
Today, I would like to show you the time series plots of several basic statistics of the bitcointalk forum.
These observed variables are:
1) New members;
2) New topics;
3) New posts.

Data
Original data
All of them have been tracked on monthly basis, since November of 2009 to December of 2017. Note that the data on months in December of 2017 is incomple, with only the first two months given by forum's stats [Stopped since 14th December of 2017].
Since that point, the forum has stopped giving away these statisitics, so I have not get later months for now.
I will try searching to find available data of those months, and update the OP later.

There is one similar topic (100 Days of Merit), published by zentdex, but it has been abandoned for months.

Converted dataset is there

Plots
1) New members and new topics

A quick overview. It seems that monthly statistics on new members and new topics tend to hit peaks each two-year period.
2) New posts

Further analysis
Will be released soon.


Notes:
The topic has still been under construction.

[1714736178]

1714736178

[1714736178]

1714736178

[1714736178]

1714736178

[1714736178]

1714736178

[1714736178]

1714736178

[tranthidung]

Dataset, after converted is here:

Code:

. list id month newtopics newposts newmems, abb(30)

     +-----------------------------------------------+
     | id     month   newtopics   newposts   newmems |
     |-----------------------------------------------|
  1. |  1   2009m11           9         37         4 |
  2. |  2   2009m12           6         44        15 |
  3. |  3    2010m1          17        102         9 |
  4. |  4    2010m2          33        394        25 |
  5. |  5    2010m3          26        258        40 |
     |-----------------------------------------------|
  6. |  6    2010m4          30        200        73 |
  7. |  7    2010m5          36        280        65 |
  8. |  8    2010m6          67        584        55 |
  9. |  9    2010m7         429       4968       370 |
 10. | 10    2010m8         303       4938       515 |
     |-----------------------------------------------|
 11. | 11    2010m9         364       2978       587 |
 12. | 12   2010m10         303       4648       315 |
 13. | 13   2010m11         394       6411       374 |
 14. | 14   2010m12         532       8840       680 |
 15. | 15    2011m1         516       8281       837 |
     |-----------------------------------------------|
 16. | 16    2011m2         902      14094      1379 |
 17. | 17    2011m3        1260      19730      2538 |
 18. | 18    2011m4        1660      24397      2924 |
 19. | 19    2011m5        4034      55331      5959 |
 20. | 20    2011m6       12279     152375     14483 |
     |-----------------------------------------------|
 21. | 21    2011m7        6697     107909      5466 |
 22. | 22    2011m8        5302      78007      3866 |
 23. | 23    2011m9        3225      57585      2452 |
 24. | 24   2011m10        2648      49434      2078 |
 25. | 25   2011m11        2002      34436      1993 |
     |-----------------------------------------------|
 26. | 26   2011m12        1962      32853      1725 |
 27. | 27    2012m1        2921      55861      2051 |
 28. | 28    2012m2        2677      51083      1991 |
 29. | 29    2012m3        2765      53534      2118 |
 30. | 30    2012m4        2504      46647      2072 |
     |-----------------------------------------------|
 31. | 31    2012m5        3692      57525      2018 |
 32. | 32    2012m6        4184      67045      2080 |
 33. | 33    2012m7        3508      68316      2109 |
 34. | 34    2012m8        4505      81790      2395 |
 35. | 35    2012m9        4265      83672      2345 |
     |-----------------------------------------------|
 36. | 36   2012m10        3854      75289      3200 |
 37. | 37   2012m11        3334      60438      2708 |
 38. | 38   2012m12        3302      57992      2766 |
 39. | 39    2013m1        4022      66163      3184 |
 40. | 40    2013m2        4512      72812      3532 |
     |-----------------------------------------------|
 41. | 41    2013m3        9614     142393      7358 |
 42. | 42    2013m4       19141     277862     18712 |
 43. | 43    2013m5       19742     344735     15512 |
 44. | 44    2013m6       14191     287307      9373 |
 45. | 45    2013m7       10814     220391      7413 |
     |-----------------------------------------------|
 46. | 46    2013m8        9919     208365      6052 |
 47. | 47    2013m9        9727     217959      5156 |
 48. | 48   2013m10        8454     180006      2927 |
 49. | 49   2013m11       19251     333103     19683 |
 50. | 50   2013m12       25109     453630     31775 |
     |-----------------------------------------------|
 51. | 51    2014m1       27070     620113     26625 |
 52. | 52    2014m2       22881     570024     25448 |
 53. | 53    2014m3       20928     573478     45800 |
 54. | 54    2014m4       15450     473110     17948 |
 55. | 55    2014m5       15091     579361     14264 |
     |-----------------------------------------------|
 56. | 56    2014m6       14200     547288     11952 |
 57. | 57    2014m7       14470     517814     11119 |
 58. | 58    2014m8       15298     489137     11443 |
 59. | 59    2014m9       14160     418902     10158 |
 60. | 60   2014m10       12092     363300      8332 |
     |-----------------------------------------------|
 61. | 61   2014m11       11053     303707      7996 |
 62. | 62   2014m12       11425     294517      8391 |
 63. | 63    2015m1       12385     317332     20794 |
 64. | 64    2015m2       11808     292022     34357 |
 65. | 65    2015m3       13216     330749     38298 |
     |-----------------------------------------------|
 66. | 66    2015m4       12021     302162      8914 |
 67. | 67    2015m5       10998     248327      7625 |
 68. | 68    2015m6       11769     257628      9310 |
 69. | 69    2015m7       12039     262282     12616 |
 70. | 70    2015m8       12163     273247     10942 |
     |-----------------------------------------------|
 71. | 71    2015m9       10468     272867      7727 |
 72. | 72   2015m10       12526     277609     68832 |
 73. | 73   2015m11       14731     269984     44058 |
 74. | 74   2015m12       22477     297934     39950 |
 75. | 75    2016m1       21854     324306     39178 |
     |-----------------------------------------------|
 76. | 76    2016m2       16622     316374     40436 |
 77. | 77    2016m3       19387     327322     24945 |
 78. | 78    2016m4       15706     334879     18493 |
 79. | 79    2016m5       16037     321836     16918 |
 80. | 80    2016m6       16815     388361     20630 |
     |-----------------------------------------------|
 81. | 81    2016m7       12678     345322     11257 |
 82. | 82    2016m8       12352     333431     11193 |
 83. | 83    2016m9       12277     311745      9923 |
 84. | 84   2016m10       12232     320979      9935 |
 85. | 85   2016m11       11595     305415     11187 |
     |-----------------------------------------------|
 86. | 86   2016m12       12179     315340     11199 |
 87. | 87    2017m1       13452     330242     13354 |
 88. | 88    2017m2       13001     327436     12677 |
 89. | 89    2017m3       17381     391524     14760 |
 90. | 90    2017m4       18200     405690     15807 |
     |-----------------------------------------------|
 91. | 91    2017m5       22759     483854     23978 |
 92. | 92    2017m6       23179     567436     32733 |
 93. | 93    2017m7       25013     654076     41011 |
 94. | 94    2017m8       25998     869899     42253 |
 95. | 95    2017m9       27470    1.0e+06     46186 |
     |-----------------------------------------------|
 96. | 96   2017m10       35572    1.4e+06     89443 |
 97. | 97   2017m11       41651    1.7e+06     97286 |
 98. | 98   2017m12       21168     752589     63756 |
     +-----------------------------------------------+

. 

Will update later if I find additional monthly data points.

[tranthidung]

Reserved for further analytical contents (not sure for now).
Finally, I have ideas for the one, which used specifically to make a tracking time-series of Weekly New posts since Sep 17 06:27 (Dutch time), at which the forum admin, Theymos announced the "Enhanced newbie restrictions & requirements" in post 45810047.
Original data used for the analysis has been available here, published by LoyceV.

Here is dataset, after converted by myself for the analysis, with the assumption that those days fit to calendar days, started from 17th Sep.

Code:

. list, abb(30)

     +----------------------------------+
     |    week   merit   weekposts   id |
     |----------------------------------|
  1. | 2018w37    5630      353813   34 |
  2. | 2018w38    7825      286408   35 |
  3. | 2018w39    4388      280503   36 |
  4. | 2018w40    4271      259694   37 |
  5. | 2018w41    3800      263685   38 |
     |----------------------------------|
  6. | 2018w42    4821      240339   39 |
  7. | 2018w43    3945      240045   40 |
  8. | 2018w44    3339      219689   41 |
  9. | 2018w45    4513      219774   42 |
 10. | 2018w46    1740      215135   43 |
     +----------------------------------+

Results presented here, in the topic made by LoyceV.

I made a graph based on data in your OP, but I think it might be better if we can have full forum post history (weekly/monthly).

Box plot

Basic statistics on central tendency of Weekly Newposts presented below:
1) Mean +/- Standard deviation: 247253 +/- 26716
2) Median (Interquartile range): 240339 (219774 - 263685).
3) Min - max: 215135 - 286408.
Compared to the median of the previous three weeks, based from @LoyceV resutls (simply asssumed it is the median to make comparision), at 353813; we can easily see the big difference.
---> The Percentage Difference Between Medians of the two period is calculated as:
(240339 - 353813)*100/353813 ~ -32.1% (nearly 32% decrease).

Time series of Weekly Newposts, since Sep 17 06:27 (Dutch time)

In a nutshell, the median of weekly newposts over the recent 9 week since Sep 17 06:27 (Dutch time) has decreased about 32% compared to the 'assumed' median of previous period.
Importanly to note about the 'assumed median' of weekly posts for the period before 17th Sep. It is assumed median because I don't have enough datapoint to get real median.

Notes:
I will keep the thread updated regularly with help from LoyceV's data dumps.

[Dig Bicks]

More pointless data that no one cares about.  Just another piece of shit fishing for merit by making some graphs.

[LFC_Bitcoin]

New users & new threads will definitely be increasing due to the recent price fall. This is unfortunately the perfect time for trolls/fudsters & genuine assholes to infect the entire forum with their absolute disease.

[Theb]

1) New members and new topics

A quick overview. It seems that monthly statistics on new members and new topics tend to hit peaks each two-year period.

I'll praised you that you gave an effort on creating a data analysis via time series but I don't it is the right way to use it. Time series is usually being use in the prices of assets to determine their movements through time. If you want to determine if their is some kind of seasonability when it comes to new members registering in the forum you should have put BTC's prices as an additional data to determine if there is some kind of connection why there are peaks in registration every two years. Simply putting two data internal to the forum won't give you as much of a reason why those peaks are happening.

[tranthidung]

More pointless data that no one cares about.  Just another piece of shit fishing for merit by making some graphs.

Hey dude,
At first, you should look at your post history, before trolling me like this.
It is too ridiculous when a guy with so many one-lined threads criticized me on how to make constructive threads/ topics.
At least, it give me a hint on potential things that can happened.
Your are right, because I can not go to any causal inference with such limited data with only three variables (all of them are independent variables), without appropriate independent variables.

I'll praised you that you gave an effort on creating a data analysis via time series but I don't it is the right way to use it. Time series is usually being use in the prices of assets to determine their movements through time. If you want to determine if their is some kind of seasonability when it comes to new members registering in the forum you should have put BTC's prices as an additional data to determine if there is some kind of connection why there are peaks in registration every two years. Simply putting two data internal to the forum won't give you as much of a reason why those peaks are happening.

It's just a assumption without real data, but you are probably right.
The important thing is the level of increasing new posts, new users, new topics recent months when Bitcoin dropped.
From the past ( as shown in the chart), three indicators tend to tell with Bitcoin drops.

New users & new threads will definitely be increasing due to the recent price fall. This is unfortunately the perfect time for trolls/fudsters & genuine assholes to infect the entire forum with their absolute disease.

[tranthidung]

Some days ago, I read a topic created by someone [who I have not remembered his / her account name] on those indicators with updated statistics till recent months of 2018.

I have intention to make a fully time series graph.

Therefore, someone read and remember where the topic is, please give me link.

Thank you for your help, all you guys.

[tranthidung]

Finally, I found it (by chance again).
Statistics of user registrations on Bitcointalk 2017-2019 (by Alex_Sr)
I will add those data into official dataset of forum data dump by theymos, then graph a new time-series plot.

[tranthidung]

I don't have data, but I believe that with the bull run of bitcoin, and past pattern, we likely have spikes for those parameters, new registered accounts, new posts, and new topics. Someone have data, please feel free to share them here, then I will give you time series plots.

[DdmrDdmr]

<...>

You can retrieve that information yourself (*1), although it may not be available at the regular intervals you want. For example, I keep track of the number of net(*2) created posts per month per Local Board/childboard (first level), and feed them to the Merit Dashboard (Post Count Evolution tab).

(*1) Official forum stats run up to December 2017 (partial month). From there on it's a manual retrival process.
(*2) Net posts created: Created posts – moved posts (threads) – deleted posts

In order to do so, I used the available information on Archive webpages (WayBackMachine, ArchieveToday) to begin with, and then make sure I add the page I need to the archieve on the date I require to capture a snapshot. From there on, it’s a manual process of retrieving the information from the achieved pages, moving it to excel, performing 2 month diffs, and getting the info over to wherever you intend to exploit it.

- Registered accounts do not necessarily need the use of achieved pages (but it can be used). Playing around with Bitcointalk profiles and performing a manual dichotomous search (or alike) to reach an ID number on a given date.

- Number of posts and threads would require archived data as indicated above.