Title: Bitcointalk user analysis Post by: mazdafunsun on April 16, 2018, 06:17:57 PM Scraped through first 1mil user IDs, end result was roughly 900K rows of users.
End result was so much smaller than IDs because my scrapper throws an exception if "Last active: Never" ('Never' isn't correct date format for a database). Quick sorted these 900K users:
So about 90% of first 1mil users accounts are Brand new or Newbie. 29K MMM Global spam accounts 4K BtcCoScammedMe* accounts 303K users last active this year (202K Brand new) this looks suspicious ??? https://docs.google.com/spreadsheets/d/e/2PACX-1vQFitTTs9-QaJtfYaMTtivpZ_ZCuyxIrPVHBfKEId7o9rvggmqHulk-DfBaz1jnTa25YonDmM1SbNNr/pubchart?oid=1036140737&format=image Title: Re: Bitcointalk user analysis Post by: user0000001 on April 16, 2018, 08:47:46 PM distribution of bounty airdrops. this leads to new registrations. people want to get more coins) nothing suspicious
money rules the world) Title: Re: Bitcointalk user analysis Post by: TheQuin on April 17, 2018, 08:31:13 AM Scraped through first 1mil user IDs, end result was roughly 900K rows of users. End result was so much smaller than IDs because my scrapper throws an exception if "Last active: Never" ('Never' isn't correct date format for a database). This is interesting but are you intending on completing it and scraping the remaining 1m accounts? 303K users last active this year (202K Brand new) this looks suspicious ??? There's a lot of spambots out there and if a moderator catches them while they are still Newbie they can 'nuke' the account. This bans them and deletes all their posts which returns them to Brand New status. Title: Re: Bitcointalk user analysis Post by: mazdafunsun on April 17, 2018, 12:42:37 PM Yes, scraping now 2nd mil user IDs. But it takes time because I limited requests/min.
Currently scraped about ~400K IDs Title: Re: Bitcointalk user analysis Post by: DdmrDdmr on April 17, 2018, 08:48:47 PM Interesting.
There are some fine stats created and published on the Meta section (some by myself). Most are related to users with a Merit Tx, but there's hardly anything on users that do no send nor receive sMerit. That could be an interesting area to cover since I think it has not been approached to give insight (I would have done it myself, but the scraper i'm using is limited in speed, so large volumes would be a hazzard, plus my free time is rather limited). The idea being to see, for example how much sMerit by rank is actually in users hands, that is not circulating (50% of aggregate Merit of users with no sMerit Tx). Or whatever you see fit... Keep us posted ;) Title: Re: Bitcointalk user analysis Post by: Vod on April 17, 2018, 09:39:02 PM Yes, scraping now 2nd mil user IDs. But it takes time because I limited requests/min. Currently scraped about ~400K IDs I think you and I are duplicating effort. :/ Title: Re: Bitcointalk user analysis Post by: mazdafunsun on April 19, 2018, 08:12:59 AM Actually, I'm willing to share scraped date (probably remove usernames from a table?).
Or you can request what kind of reports you would like to see. Title: Re: Bitcointalk user analysis Post by: fernandoaleixo on April 20, 2018, 01:20:15 AM Interesting, good to know I'm already a small portion of the userbase, following your graphic. ;D I think most brand new users are people looking out for opportunities or just register to answer a topic they're interested.
|