Better abilities for Podmins for Spam Analytics and Controls

After running my Pod for 4 Weeks, I (unsurprisingly) got a bunch of Posts from… let’s say ‘unwelcome’ posters.

As a user I can easily mark the posts to be ignored. (even as podmin), but as a podmi I have no overview about this.

On database level I can run this query:
select author_id, people.diaspora_handle, count(*) from posts left join people On posts.author_id = people.id group by author_id, people.diaspora_handle order by count(*) desc limit 100;
Which results in this:

 author_id |          diaspora_handle           | count 
-----------+------------------------------------+-------
       977 | science_bot@federatica.space       |  1048
        68 | wurstaufbrot@pod.geraspora.de      |   410
        37 | johndoe@friendica.eskimo.com       |   368
       126 | birne@diaspora.psyco.fr            |   312
        12 | rca@share.naturalnews.com          |   248
        10 | debyjolley@diasp.org               |   196
       202 | simona@pod.geraspora.de            |   185
        22 | voxpopuli@diasp.org                |   163
       356 | anonymiss@despora.de               |   127
        17 | actro@pluspora.com                 |   118
      1276 | postillon24@loma.ml                |   110
        42 | breitbart@social.miwatchdogs.com   |   101
       538 | maehringen@pod.geraspora.de        |    94
       514 | golem@squeet.me                    |    75
        35 | dapor2000@pod.dapor.net            |    71
       689 | cass_m@pluspora.com                |    68
       138 | gehrke_test@libranet.de            |    60

As a podin I’d love to quickly check what’s going on there within the diaspora Admin UI.
Next I’d want to limit spreading these posts on my pod: (“My Pod, my living room!”)

First I would like to limit its visibility ion the public stream.
Next I would like to limit these accounts to be on my pod, by deleting it.

For the latter, I now there is somewhere (at least one) a Delete Spam Script I can run from console…

What thinks the community to add such an analyzing tool and such function directly in the UI?

Why would you want to block external users when your pod is federated? If you only want to have ‘local’ posters, then diaspora is probably not your platform. Use a forum or closed chat channel instead.

I think you’ve misread the proposal. This is about giving admins an overview of ‘problem’ accounts – mainly spammers – so the admin can decide whether or not to block those accounts at an individual level.

1 Like

Thanks for clarifying. I agree that more options for podmins would be a good thing. Although if a person that has an account on a pod, and the podmin of his pod decides to block an account he is following… that may be less desirable. On the other hand: a pod is not a democracy. If someone doesn’t like the pod(min), he can find another pod or start one himself.
Also more statistics of the pod would be greatly appreciated.

Thats exactly what I have in mind. Thank you @goob.
A List of ‘most frequent posters’ as well as ‘most frequently commenters’ would give a podmin an indication where to have an eye on it.
This might help to generate a ‘baseline’ to detect outliners over time.

@tclaus, was it you who also suggested giving podmins access to a list of how many times a user’s posts had been reported? That also seems like a good idea, to keep track of troublesome users. (A parallel list of users who have made a lot of reports might be useful as well, to keep track of people who make a lot of unnecessary and perhaps malicious reports.)

I realise this is slightly off-topic for this discussion, but it reminded me about it and I thought it was an idea worth remembering.

2 Likes

A PR is handed in on this topic.

My posts may be relevant:

Not all of these users are spammers, for goodnessakes. Some of them are simply prolific posters, and friends of mine. What users debyjolly and johndoe from that list have in common is that they are politically conservative. But not spammers. They reshare a lot of stuff from other users and news sites, but that isn’t spam. So you would probably have to “filter” by #hashtags and keywords to find actual spammers, searching for words like “sale” and “limited time only” or whatever.

1 Like

This issue by OP is relevant for anyone who’s considering work on this Feature request: An 'alert' system to detect keywords from unwanted posts · Issue #8259 · diaspora/diaspora · GitHub