Datasift stream API compare to Twitter Stream API?


#1

Hi, we have a huge amount of keywords to search against Twitter. We’ ve already had a small App to get data from Twitter stream using those keywords and we get about 15 tweets/ second. We have just finished another test app, using the same keywords but search against Datasift, We also got about 15 tweets/second.

For now, We’re not caring much about enriched information like Klout, Sentiment or Trend, so for our case, could you tell us the good reason to use Datasift instead of Twitter Stream API because personally we think Twitter’s nearly free of use. To use Datasift, I have to consume a cost of 300 DPU which we hardly afford.

Thanks


#2

 

Hi,
Depending upon which features of the platform you utilize to meet your use cases will obviously define how useful the platform is to you. If you are doing simple Twitter keyword searches currently (and not hitting Twitter’s API limits) and are happy with the results, there is little benefit to migrating. If however you want to start to look deeper, and perhaps at other sources, there are significant benefits to be gained. A small sample:
 
Access to multiple data sources from a single platform e.g. Facebook, Digg, Twitter etc
Much higher daily API limits
Augmentations – As well as those you mentioned, the Links augmentation allows you to filter against fully resolved links (domains and/or URL character matches), and the title of the resulting page in real-time.
Ability to record real-time data streams
 
There is also a nice article on the subject here - http://dev.datasift.com/blog/migrating-from-the-twitter-streaming-api
 
300DPU does sound expensive. If you can email your username and stream id to support, I will happily investigate.
 
Regards, Ben.