BUSINESS

Opinion gathering: making sense of unstructured data



A lot of gainful insights can be extracted by exploring databases



Tweetfeel is yet another search service that monitors sentiments expressed by people in Twitter conversations.



Unstructured/opinion-rich data are growing fast across the Net. NetSpeak discusses this phenomenon of qualitative data explosion and the emerging services meant for managing them.

More of what we do in our life is getting captured digitally. Your on-line purchase history, e-mail data, blog postings, tweets and the like are being stored somewhere on the Net.

An interesting aspect of this data generation is that most of the data are unstructured and are available in natural language. A major share of this data contains the opinions/views of netizens on a variety of things that include products, services, people, concepts and the like. This is especially true with data emanating from social networking/blogging/micro-blogging services.

Another significant aspect of this data explosion phenomenon relates to the opening up of this data by services like Twitter. This makes the Net’s unstructured data territory quite resourceful and an exciting domain for opinion data analysts, on-line entrepreneurs, marketers and potential customers.

A lot of gainful insights can be extracted by exploring these databases. For instance, before making a purchase, one can dig this database and elicit a summary of opinions by others on the product.

These days, while making a purchase decision, we generally seek the service of a search engine.

The search may display a few web pages that contain the name of the product being investigated. But this may not be sufficient — as the output is unlikely to contain enough information to facilitate the purchase.

Instead of a few random web pages with some information on the product embedded in them, we prefer the output to be packed with purchase advices from customer reviews.

This will enable us to leverage on the experience of the people who bought this product in the past. So, we need a search service that understands the semantics of a web page- not the one that simply churns out a set of web page links.

This is the context in which the real-time search service TipTop (http://feeltiptop. com/) chips in.

As its CEO Shyam Kapur claims, the TipTop’s engine can read the data almost like a human being. The service (mainly based on Twitter data now) collects the tweets against a query and extracts (from each of the tweets) components such as entity (what is being talked about: movie, laptop and the like), attribute (what feature of the entity is being talked about) and the sentiment expressed about this attribute.

Once the data are structured in this manner, the service aggregates it, computes and displays the percentage of positive (Tips) and negative (Pits) sentiments.

So, if you are planning to purchase a product and wish to obtain the data regarding others’ opinion on the product, just enter its name in TipTop’s input-box. Tiptop collects the relevant tweets and displays them in three separate columns — Tips (tweets with positive opinion), Pits (negative tips) and general tweets.

Basically, TipTop provides you useful tips (Tip part of it) for an issue in the top of your mind (Top part of it). So, TipTop offers ‘tips’ to resolve the issue on the ‘top’ of your mind.

Tweetfeel (http://www. tweetfeel.com/) is yet another search service that monitors sentiments expressed by people in Twitter conversations. In response to your search term, the service simply displays the percentage of positive and negative views in separate columns.

Of course, the results from these services are by no means conclusive. But the concept is innovative and quite relevant.

We witness the launch of a new search engine based on Twitter almost every day. Of course, this trend may not last long as major search engines are vying with each other to have a deal with Twitter to access its database.

Anyway, till they come out with a proper real-time feature, let us enjoy the new Twitter based search services. In this regard, NetSpeak wishes to indicate yet another search service called Twoquick (http://www.twoquick. com/).

The advantage of this service is that it offers results from both Google and Twitter side-by-side (on the same page).

Video search engine

It is likely that at times you may wish to watch videos without audio. Of course, to comprehend a video film without audio you need sub-titles. Now, if you wish to find videos with subtitles relevant to your requirements, check out the video search engine 22frame (http://22frames. com/) that indexes videos with subtitles from different video hosting sources.

It is always good to know how a subject or phenomenon got evolved. Computer viruses scare most of the computer users and if you wish to know the historical antecedents of this menace, then see: http://www. wired.com/thisdayintech/ 2009/11/1110fred-cohen-first-computer-virus/. And, if you are keen to know the historical events that preceded the present day Net, read the article ‘History of Internet in a nutshell” (http://sixrevisions.com/resources/the-history-of-the-internet-in-a-nutshell/).

MS answer service

Nowadays, it seems, MicroSoft is rather keen to keep their customers in good humour by providing different means to help them out. If you use any of the MicroSoft products, you may find the MicroSoft Answer service (http://answers.microsoft. com/en-us/default.aspx) very helpful.

Even if you don’t have any questions to ask, just browse through the answer database, it might fetch you rich dividends.

He can be contacted at: >jmurali@gmail.com

J. MURALI

Recommended for you