Author

Topic: Analysis – Net Post evolution in all Local Boards (Read 340 times)

legendary
Activity: 2296
Merit: 10753
There are lies, damned lies and statistics. MTwain
<…>
Instead of “percent difference” a more useful stast would be “percent change from last month” (the difference between the net number of new posts this month from last momth. Also useful would be percent chance from last year. As boards have more and more posts, each additional 1,000 new net posts will have a smaller impact on the percentage of new posts.
Could be interesting, but quite a bit more work…
That’s why the Share part of the charts is interesting. The share shows how much weight the posts in a specific child board have in relation to the total number of posts within the analysed local board. That can be seen as an aggregate for the last 18 months (shown by default), but on the Merit Dashboard you can play around and filter by month to see the share (weight) of a child board for a given month. Doing this over several individual months gives you an idea of the evolution in share (but it is also subject to the influence of deleted/moved posts).

The following charts shows:
-   The number of historical Bitcointalk posts created between 2 different consecutive readings (*)
-   BTC price evolution in USD (source: /www.barchart.com).
<…>
 
copper member
Activity: 2870
Merit: 2298
Interesting information.

A couple of things...
Instead of “percent difference” a more useful stast would be “percent change from last month” (the difference between the net number of new posts this month from last momth. Also useful would be percent chance from last year. As boards have more and more posts, each additional 1,000 new net posts will have a smaller impact on the percentage of new posts.

For comparison, it would be useful to also see net new posts in the English sections, possibly split up between the main boards and the altcoin boards. In general, interest in bitcoin rises and falls with prices and with prices way down from the highs in 2017, interest has also fallen, and I would expect the number of new posts to have declined from December 2017.
legendary
Activity: 2296
Merit: 10753
There are lies, damned lies and statistics. MTwain
Update 29/03/2019:

Included the data for all 21 Local Boards:

- French, German, Philippine, Russian and Spanish were the original ones, and are depicted in further detail in the OP.

- Arabic, Chinese, Croatian, Dutch, Greek, Hebrew, India, Indonesian, Italian, Japanese, Korean, Polish, Portuguese, Romanian, Scandinavian and Turkish have been added to the Merit Dashboard (go to the last table labelled "Post Count Evolution").

I have not created an entry in the OP for these additional 16 Local boards (that’s what the Dashboard is for), but if anyone wants it for a specific local board, I can capture the screen and comment it here.
legendary
Activity: 938
Merit: 2540
<>
<...>

Excellent work, thanks for your reports.
I think the local sections are also very important for the Bitcointalk forum. Wink
staff
Activity: 2366
Merit: 2013
I find your lack of faith in Bitcoin disturbing.
Thanks for these stats, I always wanted to have this kind of information for the French section.

2.  French Local board.

The top line chart ("Total PostDiff") shows how the number of net post creation has evolved over the last 18 months, going from 10.991 net posts created in the month running up to the 25/11/2017 to the 1.435 net posts created over the month leading up to the 26/03/2019.
That is a hell of a drop, and Altcoins seem to have a lot to do with it (as can be seen on the lower chart).

Indeed, a significant drop. I think the main reason is that there are almost no more active "ANN translators" in the Fr section. In the high period, in one day, you could see a whole page of unread topics. Today, there are only 1 or 2 topics.

Quote
For example, if I select the columns "25/01/2019", "23/02/2019" and "26/03/2019", then Altcoins is down to 25,85% of net post share, and has been overtaken by Hors-sujet (32,30% off-topic).
Thanks Gilets Jaunes Smiley
legendary
Activity: 2296
Merit: 10753
There are lies, damned lies and statistics. MTwain
Update 29/03/2019: Included the data for all 21 Local Boards:

- French, German, Philippine, Russian and Spanish were the original ones, and are depicted in further detail in the OP.

- Arabic, Chinese, Croatian, Dutch, Greek, Hebrew, India, Indonesian, Italian, Japanese, Korean, Polish, Portuguese, Romanian, Scandinavian and Turkish have been added to the Merit Dashboard.

I have not created an entry in the OP for these additional 16 Local boards (that’s what the Dashboard is for), but if anyone wants it for a specific local board, I can capture the screen and comment it here.


1.  Introduction.

Recently, my local board (Spanish) seems to be a bit rather scarce of posts. There are a bunch of decent posters there that keep the local board beating (most of them also post in the general English sections), but engaging and building activity there is painstaking.

That motivated me to take a look into post evolution on my local board (see Análisis - Número de posts en el foro Español), and I thought that it would be interesting to do for other Local boards. Now creating a specific post thread for each local board is too cumbersome, so I’ve tried to compromise a solution (less information that a static analysis) that can be viewed on the  Merit Dashboard (go to the last table labelled "Post Count Evolution"), and extended further if I have time.

I’ve currently drawn up the data for the following local boards: French, German, Philippine, Russian and Spanish, over an 18 month period (to see recent tendency, and draw a line somewhere at retrieving information manually).


In summary, what I’ve done is the following:

- Retrieved total post count for each depicted Local Board at regular time intervals, close to being one month apart on each reading. The dates I’ve chosen are not end of month nor beginning of month. That is due to the fact that there is more snapshots available on the WayBackMachine around the 25th of each month. The actual date selection is not that essential, since I’m going to focus on Net Post creation in readings that are as close to being one month apart as possible.
The data retrieved is at local board level + first level of child boards. This tasks is manual, number by number, and I may have mistyped a number or two in the process (although I did my best to avoid doing so). It’s an interim sort of task, boring, tedious and repetitive.

- I then calculated the number of net posts created between two consecutive readings, and that is the data that shows on the Dashboard. That is, I’m focusing not on how many posts exist at a given time, but the number created roughly in a month.

It’s important to emphasize that I’m looking at net posts created, and not number of posts created.
Net posts = number of created posts – number of deleted posts +/- number of moved posts.

When you compare the number of posts between two readings, what we are really seeing is the number of Net posts created, as we cannot retrieve (without inner data files) the number of created posts on its own.


The results are interesting, but you need to be careful interpreting the data, and especially with the outlier net posts created at times:

-   Sometimes, a certain month’s reading for a specific child board is negative. I can only assume that the reason is that more posts have been deleted/moved than created. On occasions, these negative net post readings are pretty large, so I assume that a heavy moderation clean-up took place.

-   The opposite can also happen, with abnormal positive net posts created. That may be due to threads being merged as I saw on the Philippine local board.

-   It could be a typo on my behalf, but I re-checked most outliers to minimize this possibility.

Having said that, a certain background under the scenes type of context is necessary to understand the outliers, but the general trend can be seen regardless.


2.  French Local board.



The top line chart ("Total PostDiff") shows how the number of net post creation has evolved over the last 18 months, going from 10.991 net posts created in the month running up to the 25/11/2017 to the 1.435 net posts created over the month leading up to the 26/03/2019.
That is a hell of a drop, and Altcoins seem to have a lot to do with it (as can be seen on the lower chart).

On the Merit Dashboard, you can play around to filter by a specific child board/s and therefore have the charts adjust to the selection (use the Subsection filter or click/multi-select from the numerical table above the charts). Adding a screen capture for each child board here would be over the top, so it’s better to play with the Dashboard.

The Recent Share pie show that Altcoins represent 49,15% of the net posts created over the last 18 months, followed by Débutants (10,90% - Beginners) and Hors-sujet (10,04% off-topic). The French local board behaves pretty smoothly, with barely any negative values in the data table, so the overall 18 month pie is likely pretty reasonable.

Even so, if you filter the data to specific monthly readings (click or multi-select the columns on the data table), the distribution ratios vary. For example, if I select the columns "25/01/2019", "23/02/2019" and "26/03/2019", then Altcoins is down to 25,85% of net post share, and has been overtaken by Hors-sujet (32,30% off-topic).


3.  German Local board.



The top line chart ("Total PostDiff") shows how the number of net post creation has evolved over the last 18 months, going from 10.918 net posts created in the month running up to the 25/11/2017 to the 2.976 net posts created over the month leading up to the 26/03/2019. The drop is less than half of the French Local Board, having a similar starting point.

The Recent Share pie show that Altcoins represent 50,11% of the net posts created over the last 18 months, followed by Trading und Spekulation (20,38%) and the Main default (9,13%).

Even so, if you filter the data to specific monthly readings (click or multi-select the columns on the data table), the distribution ratios may vary. For example, if I select the columns "25/01/2019", "23/02/2019" and "26/03/2019", then Altcoins stays at 50,50% of net post share, followed by Trading und Spekulation (22,12%). These values are very similar to the overall 18 month values.


4.  Philippine Local board.



The Phillipine local board is rather simple, with a Main board and a single first level Child board (Altcoins).

The top line chart ("Total PostDiff") shows how the number of net post creation has evolved over the last 18 months. This case has some outliers:
-   25/11/2017: -34.454 net posts in a month running up to the date.
-   27/12/2017:  -9.354 net posts on the Main board in a month running up to the date, and +20.522 created on the Altcoin board.
What seems to have happened is a purge of the main board, and the creation of the Altcoin, largely from threads moved from the main board.

-   25/01/2019: A boost on the Altcoins of 25.631 net posts. I cannot conceive how that is possible, unless there was another section that was merged. A massive outlier, just around the time the local board was rebranded from Philippines to Pilipinas.

The Recent Share pie is messed up with the outliers, so one should really narrow down the selection to exclude the outlier months. Doing so (excluding 25/11/2017, 27/12/2017 and 25/01/2019) the share is of 70,69% of net posts created on the Main board, and 29,31% in Altcoins.



5.  Russian Local board.



The top line chart ("Total PostDiff") shows how the number of net post creation has evolved over the last 18 months, going from 219.324 net posts created in the month running up to the 25/11/2017 to the 28.699 net posts created over the month leading up to the 26/03/2019. The maximum was 354.315 net posts on the month running up to the 24/04/2018, and the minimum was -11.916 on the reading of the 26/07/2018 (likely a large purge).


The Recent Share pie show that Altcoins represent 63,51% of the net posts created over the last 18 months, followed by Paзнoe (14,36% - Miscellanea) and Tpeйдepы (5,15% Trading).

Even so, if you filter the data to specific monthly readings (click or multi-select the columns on the data table), the distribution ratios vary. For example, if I select the columns "25/11/2018", "26/12/2018" and "25/01/2019", then Altcoins represents 79,05% of net post share. There are quite a few negative value readings, so I figure that there is constant purging of threads.


6.  Spanish Local board.



On this local board I have already excluded the data for the month leading up to the reading of the 26/03/2019. This is a massive outlier, with net negative large numbers for posts on the Altcoin and Market and Economy boards (I assume some massive clean-up has taken place). I’ve also excluded the reading of the 25/11/2017, since there were 1.876 net negative posts in the Beginners and Help child board (Primeros pasos y Ayuda).

The top line chart ("Total PostDiff") shows how the number of net post creation has evolved over the last 18 months, going from 6.032 net posts created in the month running up to the 27/12/2017, to the 842 net posts created over the month leading up to the 23/02/2019. The post number decline is more pronounced from December 2018 onwards.

The Recent Share pie show that Altcoins represent 71,70% of the net posts created over the last 18 months, followed by Mercado y Economía (11,57% Market and Economy) and Principal (7,17% Main board).
Jump to: