Sentiment analysisWhat does sentiment analysis mean?

Sentiment analysis to put it simply, “I got a great deal for game XYZ at X Store!” would count toward positive opinion, while, “X Store just ripped me off for the last time!” would go in the negative column. A neutral comment might be, “An X Store just has new offers for the game XYZ. I think I’ll go check it out on my lunch break.” However, it’s not always that simplistic. There are many shades and levels of sentiment to consider, and a single article, news, tweet or post can contain several.

A sentiment analysis involves evaluating online opinions based on specific words. The sentiment is then judged to be positive, negative or neutral. If you are interested in more background information, you should visit the Wikipedia article on this topic.


Built-in sentiment analysis

We showed a prototype of our sentiment analysis on the dmexco exhibition in Cologne in mid of September 2017 and the feedback was just great. Now a few weeks later, after spending some time in the machine room, we are proud to say: Today we published our latest API release v2.2 to the live environment. One significant enhancement: the built-in sentiment analysis! No extra access or costs – it is included in our standard response of the system.

The hyScore API service now determines on the fly the sentiment of a URL/website, a sentence, a comment or a complete article sent as text. The sentiment is expressed in the JSON response as a value between +1 (positive) and -1 (negative). A “0” zero is usually neutral. You can define by yourself which sentiment value express which opinion.

Sentiment analysis example

You could use three opinion categories like in the example picture above or slice it into more,… it is up to you. That is useful for a couple of new use cases where you can use the hyScore API.

Some use case ideas:

  • You can send the comments of your social media channels (e.g., Facebook) or the comments of a product in a shop as text towards the API and you’ll see if the opinion and mood are positive, neutral or negative.
  • In combination with the contextual analysis, you can do a more granular and detailed environmental assessment regarding Brand Safety. You get for example by analyzing a website the keywords “Spain, Mallorca, Calla Millor, Rent a Finca, Swimming pool” and the site/IAB category “Travel.” But! Is it a positive or a negative written article? Maybe the article says summarized: “Great island. Great swimming pools, but the Finca’s in Calla Millor are mostly horrific.” Do you want to advertise a travel agency offering these specific Finca’s in this environment? Questionable.
  • You can create an automated active monitoring of your brand on any website and see what is written in the context of your brand and is it more positive or negative.

… and much more.

You can imagine that the sentiment analysis, combined with our contextual data is adding another level of transparency and provides even more insights. Do you want to make a trial? No problem. Get your free API-Key now!