Twitter developer Arne Roomann-Kurrik has announced on theTwitter blog that the social media micro-messaging service is introducing new metadata to tweets in order to help developers “work with targeted subsets of Tweet collections”.The improvements will help Twitter app developers filter and organise tweets by the language that they were written in an identify tweets that Twitter considers “high” value.
The language that the tweet has been written in is identified by Twitter’s machine language detection algorithms and marked by a new lang attribute in the metadata. “The values will be valid BCP 47 language identifiers,” says Roomann-Kurrik, “and may represent any of the languages listed on Twitter's advanced search page, or "und" if no language could be detected.”
This language update allows systems that compile and organise Twitter data, “such as analytics services or real-time search streams, to offer language-specific curation, aggregation and analysis of Tweet content.”
The filter_level attribute values Tweets from a stream as either "none", "low", or "medium", with a final classification, "high", to be introduced later. “The "medium" (and eventually "high") entries will roughly correlate to the "Top Tweets" results for searches on twitter.com”. This classification system will make it easier for apps to surface high-value tweets and other content from “noisy” feeds.
Both filter_level and lang will be available on Streaming API responses, with the latter also on REST.filter_level will be activated first, on Wednesday next week, with lang coming shortly after. However, the release times are subject to change, so Roomann-Kurrik advises that developers keep an eye on the calendar of API changes.
Here is an example of the new attributes present in metadata:
In the example above, the tweet is given a classification of "medium", meaning it is a quite high-value result in the search, and the language is English, as specified by “en”.
Public streaming endpoints will also support the filter_level and lang parameters, providing filtering of streamed data on Twitter’s side of the process. Users can connect to a public stream and specify the languages of the tweets that their search will return, as well as setting a filter_level parameter.
At some point, Twitter will undoubtedly introduce the changes into the Discovery tab. However, there is the possibility that, instead of being organised chronologically, Twitter’s default home stream could be organised in order of value to the user.
Would you like to see your Twitter stream organised by importance?
0 comments:
Post a Comment