While your writing...

What it does

During the Media Hackday in Berlin, we have created a tool for content editors to use. While the editor is writing a story in the tool, he will get feedback about the story that he is writing. For instance:

It will give the editor a feel about how positive or negative the story is.
It will return an overview of other articles about the same topic. That way the editor is able to prevent writing duplicate content.
It will suggest video's about the topic that the editor can refer to in his story.

How we built it

We build the application in a React front-end and it's communicating with a Node back-end. Every time the editor is pressing the "Enter" key, an array of paragraphs is sent to the Node back-end that triggers the Cloud Natural Language API and interprets the content. The API returns the most relevant entities with the highest salience score. It also returns the sentiment analysis of the content (Positive, Negative, Neutral, or Mixed). Now, the editor has the option to do something with this relevant information. We used the iPool API for doing a social search (Twitter in this case) and we're used the iPool API for listing articles on the same topic. And we're using the YouTube API to find video's on the same topic.

Difficulties

In the first concept of the tool, we used the AWS Comprehend API. Unfortunately, we got very frustrated to see that during the hacking, the AWS account got hacked by someone and lost a lot of the work.

But, we're happy to found a way to use the Google API without entering Credit Card information. So we used the Cloud Natural Language API and everything works fine now.

API's used during the Media Hackday in Berlin