Features

What makes BuzzTalk unique?

World wide monitoring of over 58.000 sources

  • BuzzTalk monitors well over 58.000 manually checked sources from over 100 different countries. It contains tweets from consumers, blogs from specialists, news from journalists and journals from scientists. The number of sources grows continuously via Crowd Sourced Learning and manual additions.

  •   

    Multi lingual analysis across countries

    BuzzTalk is unique in that it can translate 67 different languages using Systran translation technology. This way you can learn how Chinese or Russian people are thinking and feeling about your brand or industry. Find relevant developments in other continents that help define your strategy.

  • Crowd Sourced Learning

    BuzzTalk enables Crowd Sourced Learning which means the system continuously enriches itself. It also allows for manual additions so you can add your own favorite sources.

  • White Box

    BuzzTalk is a fully transparent “White Box”: it allows you to retrieve the original documents; it can also display their sources and show you how the software analyzed the sentiments.

  • Extensive tagging of publications

    Media analysis is only possible when publications are tagged. BuzzTalk is a very powerfull tagging engine which enables a variety of analysis.

Available tags for media analysis

Using tags it’s easy to extract the who, what, where and when from a large list of publications. But there’s more.

  • Bare sentiment and ontology based sentiment

    BuzzTalk can determine the sentiment of tweets, posts and articles on websites, blog and forums. It can do this with 90% accuracy, which is very high. This enables you to track sentiments across time for your brand.

    You can view the bare sentiment as a calculated average of sentiment detected within an article. Also, you can view the ontology based sentiment which is the sentiment in relation to a particular subject.

     

  • Mood State analysis

    Using our sentiment tagging engine OpenDover, BuzzTalk can detect different mood states in blogposts and articles. These are: vigor, depression, confusion, fatigue, tension and anger. Mood states can have a predictive value as described by Bollenet al.

  • Domain detection

    OpenDover also detects 27 different domains and tags the articles accordingly. Examples of domains are: economics, politics, law and disaster.

  • Economic activity detection

    BuzzTalk recognizes 170 different economic activities which can be used to filter results for business analysis. Examples are Automotive industry, Music industry, Insurance industry and Recycling industry.

  • Category detection

    Using Reuters’ fact tagging engine OpenCalais, BuzzTalk can detect 36 different categories in blogposts and articles. Examples are person, organization, industryterm, marketindex, etc.

  • Spikes detection

    BuzzTalk recognizes relevant peaks in the buzz, which are deviations of OpenCalais tags in the moving average.