I just came across Radian6, a start-up in Fredericton that is focused on the analysis of online social media:
Radian6’s core technology, SentimentLive, is designed to identify, correlate and deliver the following elements from Social Media:
Topic
A topic represents the key concept being tracked such as a brand, product name, movie, sports team, politician, country, or celebrity. SentimentLive applies advanced proprietary analytical techniques that go beyond basic keyword searches to identify topics with minimal false positives or spam. Social media is gathered, indexed and presented to the user in summary form as it is posted to the Internet, in real-time.Sentiment
As topics are identified, the system derives the sentiment around the instance of the topic. By applying advanced sentiment analysis, users are able to see trends in attitudes and opinions relating to their topics of interest, which in turn enables them to make more informed decisions.Influence
As SentimentLive identifies topics, it also calculates the relative level of influence associated with each posting and with each consumer that posted about a specific topic. This enables organizations to apply resources to areas that will have the most influence on their business – either positively or negatively – and also provides a view into the top-most influential consumers who yield the strongest word-of-mouth (WOM) influence over specific topics.
In typical Dot Com fashion, there is not much more information on the website, and no evidence of any two-way web tools to allow the company to interact with its market. In reading what is posted on the website, I wonder how this offering differs from free tools such as Technorati, Alexa, TagCloud, BlogFlux or dozens of other web applications that anyone can use to monitor the Internet buzz.
There’s always room here to post a comment ;-)
I feel like yelling “been there, done that”.
This is old. Sooo old.
However. Before you invest in this company. You have to realize that analyzing the sentiment reflected in a text is no trivial matter. People get English degrees while studying this topic.
There is no way in h*ll software can do anything but some probabilistic tagging which will be often wrong.
If you think this might still be useful, you have to watch out. Basing decisions on hard data is already difficult enough (should I invest in my portfolio or pay my house?), I don’t think we are nearly sophisticated enough to do anything useful with sentiment analysis.
Of course, please, prove me wrong.
Oh! And it is a NRC-hosted company (by the address).
scary stuff. Lovers, husbands and wives, children in school, bus drivers and transit users without correct change, all kinds of people daily can’t sniff out sentiment and yet this software does it with pixels! My aunt told a story of a raging rouge (liberal) in Montreal in the early nineteen hundreds who had his vote cancelled for 30 years by his tiny, smiling wife. So much for sentiment.
Daniel, I wasn’t planning on investing in this company.
I could use a little advanced sentiment analysis. You guys really need to lighten up. Are you suggesting that, for instance, if this were reading fark, it wouldn’t get a decent read on our society?
sentimentlive.
I think I’ll call them.