Skip to content

mbellotti/netback

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

5 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Netback

A tool for finding suspicious Twitter accounts that are connected to yours

Installation

Netback is written in Go, so follow these instructions to make sure Go is installed on your machine. Then just run the following command:

go get github.com/mbellotti/netback
go install github.com/mbellotti/netback

You can also download it from source and execute the program directly from its directory if you prefer. Commands in that case would be go run main.go [command] -flags instead of netback [command] -flags

Configuration

Netback needs Twitter API access through your account. The program does not need anything more than read permission on public Twitter data. You'll need a developer account on Twitter. Create a new app and copy the API key, API Secret, access token and token secret. Then run netback config and enter in that information.

Setting up a Profile

The bot profile is the criteria Netback uses to determine whether an account looks suspicious or not. You can configure the profile to check for low activity accounts that might be fake followers or high activity accounts or some custom combination of features from both. Netback will ask you to set a weight for each criteria, assigning a weight of 0 effectively turns that criteria off.

The current possible criteria to use include:

  • Posts per Day: calculated as total posts over days the account has existed.
  • Number of Followers: with direction, meaning you can tell netback that anything over 100,000 followers (+100000) is suspicious or anything under 5 followers (-5) is suspicious.
  • Percentage of Unoriginal Posts: Retweets without any original commentary.
  • Profile Pic: Do they have one?
  • Low Posts: Are their total number of posts too low?

Netback includes two flags --low-activity and --high-activity that act like short cuts, zeroing out criteria not relevant and only prompting for a subsection.

Setting a Score

At the end of the profile wizard, Netback will prompt you to set a score. This determines the number of points an account needs to be labeled suspicious. For example if your profile gives percentage of unoriginal posts a weight of 3 and Number of Followers a weight of 2 then a logical bot score would be 5. With more criteria, setting up a bot score lower than the maximum value of all weights allows for some flexibility in how accounts are assessed.

Testing your Profile

After building a profile and setting a score it's a good idea to test Netback's analysis against an account that you know to be legitimate-- your own. Run netback self will compare your own account against the profile you've built and provide feedback.

Rate Limits

Twitter API only allows for a certain number of requests in a 15 minute period. When the rate limit is hit, Netback will alert the user, save it's current cursor position and sleep for fifteen minutes. Because the cursor position is saved, you can terminate the program while it's sleeping and restart it later on.

Followers -vs- Friends

Twitter defines followers as people who follow you, and friends as people you follow. Netback will search for suspicious account in either list.

Verbose Mode

netback self netback friends and netback followers can all be run in verbose mode which will output the observed values for every criteria with the allowed values set in the profile. Occassionally Netback will return a blank user. Typically this means the account has been suspended. It is still listed as a friend/follower but is otherwise inaccessible.

About

Looking for suspicious Twitter accounts connected to your Twitter account

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages