#CognitiveServices – Text Analytics API, new operation: detect key topics in documents

Hi !

Today’s review is for the final service included in Cognitive Services Text Analytics Api: Detect Key Topics.

This is a new service which returns the topics which have been detected in multiple text articles. The service is designed to work well for short, human written text such as reviews and user feedback, and can help you to understand the main issues or suggestions that customers are mentioning.

This service is kind of different from previous ones. Let me summarize biggest differences

– For best performance, limit each document to a short, human written text paragraph such as review, conversation or user feedback.

– You need to send 100 documents or more to start the key topics detection process. If you don’t sent this, you’ll get this nice message

Clipboard02

– Every time a document is processed, a transaction is added to your account. So be careful with the testing phase, you can get to the free limit very soon.

– There is no sync response from this service. After you sent the document collection, you’ll receive a JobId, and you need to query for the status of this JobId

{
    "odata.metadata":"<url>",
    "JobId":"<JobId>"
}

After you query for the job status, you can get a result similar to this one

{
    "odata.metadata":"<url>",
    "Status":"Running",
    "TopicInfo":[],
    "TopicAssignment":[],
    "Errors":[]
}

So, as a final comment. I won’t add a complete sample of this into the Console App, just to be sure about my testing limits. The contracts and operations are added into the Client Library project, and feel free to use it and share your experience.

You can download and use the sample from the GitHub source Code.

GitHub Sample https://github.com/elbruno/Blog/tree/master/CognitiveServices

Greetings @ Toronto

-El Bruno

References

Leave a comment

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: