Naumi – I’ve always been a big fan of connecting visualizations for data. In fact, I’d like to see more device-delivered examples. Everything from that mood globe that one could set on their desk indicating the sentiment of traffic related to alerts/posts on their keywords. All the way to visualizations of movie review sentiments, that would tell you at quick glance more than thumbs up/down, but provide a sense about if it will more likely fit one’s tastes (quirky, dark, cynical, feel-good, etc.)…delivered via an iPhone (or smartphone)display, of course.

[...] [...]

Interesting post! We’ve just launched our sentiment analysis tool that captures facial emotion for market research and mood tracking. We plan for a geographic mapping of status in real time, a kind of distributed mood ring. To your list of enterprise applications, we’d add health tracking and early predictors of market shift.

[...] Listening platforms and sentiment analysis tools allow companies to capture customer preferences, complaints, feedback, and queries expressed [...]

Very revealing article! I totally agree that we will be “blessed” or “confronted” with a new virtual layer of information, and people indeed can transform this new information in many new insights, creating a new level of understanding.

Somewhat further in your article, you mention , between brackets, : (and complexity). And this complexity is what is interesting me:

Would it be possible to exchange ideas on this ? I would like to understand what you mean by “complexity” in this context. I would say that it adds to the number of “non-random” relationships that can occur between people using this new layer of information. So the level of “organized” complexity increases. And if the theory is correct, we should then witness an increase of emergent phenomena. The only thing is that I have not yet come across examples of this new ”feeling information based” emergent phenomena. Perhaps you have already spot some examples.

The reason for my interest comes from a conference “Complexity: Friend or Foe”,that I am organizing December 16th in Brussels, Europe. The seminar will focus on two topics : “complexity and mechatronics” and “complexity and wikinomics”. We (the Flemish Society of Engineers, http://www.kviv.be/unidentified/over/kvivinenglish.aspx) have been able to attrack a big name as key note speaker : Geoff West, chairman of the Santa Fé institute for complexity research.

Perhaps there could be an opportunity to involve your “charting emotions” case in the conference. Since we have to freeze the conference set-up soon, it would be nice if you could contact me quickly. You can reach me by mail or skype (erik.van.den.broecke)

Regards,
Erik

Hi Erik,

Thanks for the interest in the research. In this instance I was thinking of complexity in relation to the amount of data and the fact that the vast majority of new data sources produce unstructured data – comments, video, voice, images, etc. So, as I mention in the post, we are privy to new levels of understanding, but in order to get this understanding we have to distill it from a many different types of raw data.

The other part of complexity is that the unit of analysis has grown considerably. When everything was local, we only cared about data from small groups, local geographies, target customers, and specific competitors. Now, the unit of analysis is the world. Everything is connected and so conducting trend analysis and identifying causality is extremely difficult.

The last thing I would say about complexity is that because the data being generated is unstructured and not tied to a predefined taxonomy, what you find is new information about virtually anything and everything imaginable. I read an article from some scientists at Google recently called “The Unreasonable Effectiveness of Data.” While much of it was over my head, what I did gather was that the requirement to seek out specific data is reduced because the corpus of existing digital data is so huge (the example in the paper was of text analytics), that we can find patterns related to whatever topic we might be interested in. This is tied to your notion of emergent phenomena. I think you’re definitely right about seeing more of this. As an example, sentiment analysis tools for customer analysis have a far greater scope than market research. In traditional market research, you have to identify the trend you are looking for, then design a question set, and finally find subjects to query. With sentiment analysis, you are mining the existing conversations of customers and prospects and using this large pool of data to identify trends (assuming your company is big enough and you have a large enough pool of customers). In this case, questions you never thought of asking will be answered and topics you never imagined will emerge.

Of course, as you noted, there are probably many more layers of complexity at play here. I think we’re just starting to scratch the surface : )

[...] would be like emotion mining – in real time, upfront and [...]