Is there any way to filter out adult content from twitter stream


#1

Dear All,

I am new to Datasift. We are trying to get some social content from twitter stream. But we some adult content/picture popping up from time to time. I am just wondering if it is possible for us/Datasift to filter out those contents.

Thanks in advance

James


#2

Unfortunately there is currently no target available to filter on adult content. We do frequently see users write their own 'curse word filters' to explicitly filter out any interactions containing curse words. You would do this like so:

twitter.text contains "my keyword" AND NOT twitter.text contains_any "curse word, rude phrase, unwanted content, etc, ..."

I have raised a feature request with our Product Team to implement the ability to filter on Twitter's 'possible_sensitive' field. This would allow you to filter out any interactions containing links to content which the content author deem to be 'potentially sensitive'. 


#3

We have now released this as a new target and output field. You can filter out Tweets and retweets that are marked as "possibly sensitive" with the following CSDL:

interaction.content contains_any "my keywords" AND

( not twitter.possibly_sensitive exists AND

  not twitter.retweet.possibly_sensitive exists )