Topic: The real activity of bitcointalk.org + full stats (Read 615 times)

I do not fully understand what you meant. Do you mean that TMANwas not right to be intrugued  by this particular work? And the reason being that there is too much information and not enough conclusions ?

You were surprisingly intrigued by the work. It actually took a lot of ingenuity to compile and draw conclusions on such a wide range of forum members.

Yes - that happens often to me. I misread it. I'm dyslexic. You should see how often I have to edit my posts for typos. 

It sems that you have misred the information.  
I did use information which was scraped in last days of Septemebr and used 2 time periods to analyse the activty:
1) Within this year.
2) Within 3 last months (24/06/2018-24/09/2018)



As you can see in the table that at the time of a scraping , there was about 7300 Jr and 28000 Members. As we already disucessed here, from the time of scraping there defintely more Jr have been made, so the number is higher than 7300 at this moment but i doubt that it would be more than 28000.

Last time I scraped the forum and made statistics was about 50 days ago  (https://bitcointalksearch.org/topic/m.45460080) and then forum had about 72000 Jr.


it depends with what we compare it with, if compare 24376 users who have a signature to 31211 users ( Jr members and up) who have made a post in last 3months it makes sense. But yes, at first glance it seems small.

How did you establish the user rank on the 3 year chart.

For instance in 2016 I was a newbie but now a hero rank.
How did you ensure that my posts were logged as newbie, jr,member, member, full member and sr.member during the last 3 years ?
How did you account for deleted posts ?

With the recent de-ranking there will naturally be more newbies. But also a lot more Jr.Members since it is easy to get 1 merit if you are not a complete shit-poster but progressively more difficult to get 10 for a member or 100 for a full member.

The figure seems a bit low though. Just looking at the Merited Receiver profiles I get a figure around 12K, so I expected a much larger overall figure.
As to the rest of the data, I find it very revealing as it gives us the insights I posted above. Thanks for the work.

There is no Magic to it. What is does require is some effort in posting, both on the content side and on the grammar and expressiveness. There is no way anyone is going to improve their rank in a legit way without these ingredients, and a hocus-pocus will no longer suffice.

Its use to surprises me , how people magically grow their
Rank, which will be my area of interest to work on it.

I assume that you meant , the difference means that there is around 143.558 active Brannd New accounts . Actually I just checked the activty by " last active" for Brand News and the number is higher and it is 156602, which gives us total activty of 364214 users by using " Last active" crteria. This number 156602 is interesting because it is huge for Brand New rank. The majority of this Brand new activty could be explained by the number of users registered in this time period and it is 90163.


I think that majority of people just have their accounts set on " never log out" and are just browsing the bitcointalk.

---

it is not possible to filter those who are posting reports, i do not have acces to bitcointalk full database , only the scraped profile information and the time of the last post of a user.
I did filter those who have a signature on across all ranks, and the number is 24376.

What about active members who are bounty hunters? If you can scrape the post history of each members, you will be able to find how many active bounty hunters are posting their facebook/twitter links on bounty sections. Add those people who are wearing ICO signatures only, if you can. Some ICO signature wearers doesn't post "proof of participation" on bounty section, they could be using google form for registration or they might be posting on service discussion section, it can be hard to trace them, but the ICO signature that they wear can also make them a bounty hunter(you can't be precise with this, there are some members who are just bounty managers wearing their ICO signatures, but they're only few). You can also sort the ranks of bounty hunters if you want.

I already asked this to some data analysts here, and it's quite hellish for them.

P.S, I can't pay you anything, I just want to see if how many active members here are bounty hunters.

And there was me saying I should stop handing out merit in Meta..  shit like this is awesome, great work bud

What seems to me most interesting:

1.   Vod’s BPIP indicated that there are 351.170 Active Profiles (logged-in in the last three months). Your data, although not synced to @Vod’s frequent update schedule of BPIP, shows us that there are 207.612 (Adding “Activity by Last Active” column) Active profiles by the same criteria, excluding Brand New.
If this were correct, that would mean that there are around 143.558 Active Newbie accounts. These would be therefore created and not posting. That is a hell of a lot (including casual readers, bots, etc.), and if correct, is an interesting fact.

2.   The two different potential criteria for calculating an active user yields 70.490 non-Newbies that log-in, but do not post over a three month period. That is just over one out of every three people that logged-in over a three month period (33,95%). These are people that act as readers, and do not engage in posting on the forum (although many could be occasional readers).

3.   Comparing last post within a 3 month window vs last post within this year, there is a difference of 176.487 users. That is therefore the amount of users that have not posted in the last three months, but have done so at some point during the current year. That represents 43,72% of the people that post, and looks to me like a large amount of dropouts from a posting point of view (many probably welcome, but even so).

4.   The same exercise as above, applied to the criteria of having logged in (regardless of having posted) within 3 or 12 months indicates that 177.881 people have not logged within the past three months, but have done so at some point this year. That is to say, 46,15% of the accounts that have logged in this year have not done so over the past three months.

Correct me if I'm wrong in any of the above..

It seems that you are right, the difference between types of activites is 207612-137122=70490 . I havent given much attention to merit analysis mostly because you are doing such a good job of it. But maybe i can look at the differences of merit between these 2 activty groups.



Hm, it  possible . I started the scraping on 24th of Septemebr , the new system was implemented to 17th of September. I think that most of the ranking up was done in the first days but some of it is ongoing, there defintely is more Jr. now but the difference most likley is not a big one. Nevertheless we will see that only when I or somone else has a fresh batch of data.

This seems rather interesting. I’ll take a proper look tomorrow if I can, but from what I’ve seen now, roughly there is a 70 k difference between the different criteria that could be used for defining an account as being active in a short time span (3 months): last active vs last post. That is quite a bit. Both criteria are valid, but if we want to compare ratios of merited users to a base of active users, then the last post criteria has more meaning, and this OP gives us a baseline.

Also I’d say that a lot of the rank changes jr. member-> newbie -> jr. member may have caught the scraper off-guard since they occurred quickly during the scraping period, so some data attributed to these ranks may have shifted (correct me if I’m wrong here).

I have successfully modified the scraper to scrape the time of the last post of any user, enabling to use this information to explore user activity on this forum more precisely . I have scraped user profile information + time of the last post up to user id : 2374859 . The scraping took place in last days of September.

In this post I gathered the complete information of activity by " last post" and compared it to activity by " last active" , the comparison showed significant differences. I did not show the " Brand News" since they have not made any posts and their contribution to forum is almost non existent. Also fresh full stats by rank can be seen in the end of post.

As before, I used last 3 month period  and also looked as stats within this year ( from 1st of January to 24th of September)

Here are the number of users who have made posts in these 2 time frames by rank:

Including Newbies , there are total of 137122 users who have made a post in last 3 months, if we exclude Newbies , the number drops to 31211. If we take a look at how many posted within this year, with Newbies there are 313609 users and without them 46149 users.

Here is the activity by " last post"  in percent:

The percent graph shows that there is very high activity from Jr.Members, this is because of the last rank requirement change- most of inactive users where demoted to Newbie level.

Further I compared " last post " activity to " last active" activity of users , this shows us that there is a significant amount of users who have been reading the forum and haven't been participating in the discussions. It was quite surprising for me , I was not expecting so many users to be reading and not posting across all the ranks, especially the high ranks. Here is the comparison graph:


Here are the difference between number of users who have posted and those have been " last active" :

Here is the data I showed above in graphs with full users statistics , which shows that right now we have a lot more Members than Jr.Members around

If you have anything you would like to see with this current data, i would happy to hear your ideas.