Streaming API missing hashtag only tweets


#1

Using the streaming API tracking a specific hashtag I am missing tweets. I have noticed a pattern in the tweets missed that each of them ONLY contains hashtags. If the user includes plain text with the hashtag(s) it comes through just fine. It is quite important that I also get the tweet even if it is just a list of hashtags.

Any suggestions or workarounds for this?

Thanks


#2

Hey Ben,

Could you provide an example Tweet (ID or link) and the query you were using to track it?


#3

Hello compston,

Example tweet that was missed.

Example tweet that was collected:

The query used to get responses was:
https://stream.twitter.com/1.1/statuses/filter.json?filter_level=none&language=en&stall_warnings=True&track=%23benstest

Thank you,
Ben Bruner


Twitter Streaming API - too slow to pull data
#4

Got it - thanks. That actually wasn’t delivered because of the ‘language=en’ parameter in your request. That only allows Tweets where the text in that specific Tweet is able to be classified as English (as opposed to the user’s profile or something else). There wasn’t anything that could be recognized confidently as English in this one, which is why it wasn’t classified, and didn’t get delivered.

If missing Tweets is a concern, keep in mind that there will never be 100% language classification coverage or accuracy, so you would always expect to miss a few if you filter on it.

Hope that helps!


#5

Worked like a charm and we seem to be collecting tweets that were previously missed.
Thanks compston for the quick follow up and detailed information.