What is the connection between big data and social media?
While Social Media as a term is well-defined and understood, BIG data on the other hand is still a contested term and often divides experts on its definition. Here is a study which asked over 40 experts to define the term, and the diversity in their responses is evident. One of the most popular definitions by IBM is that BIG Data is defined with 4Vs: Volume, Velocity, Variety, and Veracity. If we were to match a data domain with this definition, social media would probably be the best match. In fact, it would lead by a large margin in the first 3Vs, i.e Volume, Velocity, and Variety over any other domain we can think of.
Why is it important?
To extract comprehensive and valuable form Social Media data, it becomes imperative to get large volumes of user opinions from a variety of sources and across multiple time periods. While it may be possible to do a small sample snapshot analysis of social media using traditional techniques but to check trends over large time windows and to validate insights across multiple sources requires use of BIG Data techniques.
Apart from extracting insights, BIG Data processing techniques are needed to improve the social media platforms as well. Let us consider the problem of recommending interesting content to users. While techniques to identify trending content are well-established, what makes the content interesting to a particular user is its relevance to him or her.
Currently, relevance is not just limited to static user profile details but to highly dynamic user activities across different platforms. BIG Data techniques are the only solution to address this part of the problem.
What are the challenges?
- Combining multiple sources: Social Media includes many different types of information and not just different types of platforms. Influencers, trending topics, sentiment, and popular websites are some primary examples. Social Media, when combined with other enterprise data sources, can present a complete 360° view.
- Getting past the real time barrier: BIG Data techniques, such as Map Reduce and their open source implementations like Hadoop, have a lot of qualities, but one area they lack is real time analysis and results. There are two approaches to overcome this barrier. The first is the obvious one to improve parallel computing architecture to reduce response time, but more importantly, we need to focus on application areas like insight extraction where lack of real time information is a lesser barrier. If we could analyse big volumes of data over a large time window, albeit non real time, that would enable us to predict response patterns for a variety of scenarios in the future.
- Discovering business value: Fetching, Storing, and Processing BIG Data volumes from Social Media solves half of the problem. Real benefit comes when we can extract business value from those. Thus, using BIG Data in a way to support overall business strategy and as a guide in answering strategic questions
If you are want to understand how social business intelligence can give you deep insights about your brand, customers, and competitors, and help you drive growth of your business, feel free to request a demo. You can also follow us on Twitter to get the best content about social media marketing from across the web.