[c4g0bs5u6m]

Hello,

I am wondering if anyone can share any resources for creating a farm of social media bots for a given service, eg Twitter. I have no experience in this area but I do have programming experience (Python, C and some web languages).

I have begun some research but have so far not found much in the way of information. I did find this subforum and various short guides outlining the process. One specific question I have is that if you have to specifically register an app to use the Twitter API, why is there discussions about which accounts are bots & the different methods of bot detection? This is especially relevant as I was reading the way in which bots are detected (such as regular timing of tweets etc) and I was interested in the problem of making them seem as human as possible.

To give an idea of what I'm looking to accomplish, I'm not all that interested in setting up a huge number of accounts or anything. I'm thinking of setting up a small group of accounts (perhaps 5, I'm not really sure) and having them interact with other accounts as well as each other but while making them seem as human as possible. I may add more accounts at a later date. It'd be nice if much of the essentials had already been done (no need to reinvent the wheel and all) and then I could simply customise them to make them less obvious as bots.

Any and all help, advice, etc is appreciated.

[Jacque Fresco]

Hi,

i have no learning materials for you, but an approach would be to have your bot "learn" to be human. no joke, im speaking of ki or deep learning.
however thats probably out of scope. another way would be with data mining. to be able to build up a bot you will have to understand how people use social media. you can do that by collecting alot of data and try to understand the inner workings of human behaviour. build up statistics, research how and why they do what and when. you can then feed all of it to a bot and have him act like a human. you can even specifically search for people on socal media that exactly do what you want your bot todo, analyze these users and build your bot around their behaviour, so that it would make sense that your bots do certain things.

i suspect these platforms do the same analyzings with all the data they collect. the more they understand how their users work internally the more they can adjust it to them and also detect bots of course.

thats essentially whats it about, to be able to have a bot being a human you first have to understand the human. so its a scientific problem and data mining seems to be a valid approach for me, as the bots imitates the behavior of other users by using the collected data, following patterns etc.

for example if it would be your goal to create a gaming bot that plays a game like a human you would find out via data mining that there are several different gamer types https://en.wikipedia.org/wiki/Bartle_tax...ayer_types
thats the same with social media users.

report your finding, im quite sure that there is already alot of research on that topic. may type a request to some university in your country regarding materials about people behaviour on social media, may someone can point you into the right direction.

Edit:

sounds complex and all, but for real all you do is that you write a data mining bot that monitors, lets say 1000 users, and saves all they do to a database. do that for a while and build up some tb of data. now start thinking about what you could learn from it, start for example by investigating when people from where are online. make statistics to find out when which people are online, so that you have a db telling you exactly all the users that have a fucked up day/night cycle. you can do that easly by reading the when they where active. see, since you then have a list of people with a good and bad day/night cycle you can go on from there and find differencies and more... i mean be creative, the topic is quite large but fun. all start is hard. but if you do it seriously, your bot will be good.