Extracting value from big data

The phenomenal growth of social media, mobile applications, sensor based technologies and the Internet of Things is generating a flood of “Big Data” and disrupting our world in many ways. This data is becoming strategically critical to enterprises of all sizes and types. Fueled by new technologies, companies are routinely generating upwards of 20 Petabytes of data each day – a petabye is one million gigabytes or approximately 6 billion digital photos or 20 million four-drawer filing cabinets filled with text.

In today’s world, it is not enough for companies to track their sales, marketing, financial, and other internally generated data. They need to combine their internal data with external sources of data such as blogs and reviews about their products, Twitter and Facebook comments, as well as data from online discussion forums, to develop insights for improving performance and remaining competitive. The challenge here is to deal with a nonstop flood of data being generated at an increasing rate.

This talk will examine the paradigm shift caused by Big Data and focus on how to use “Data Science” to harness its power and create a smarter world. Much of the discussion on Big Data has centered around four “Vs” i.e. Volume, Velocity, Variety, and Veracity. This talk will delve deep into several other interesting and important characteristics to understand the nature of Big Data. These characteristics make it challenging to model and manage big data, yet, they provide the potential to unlock the value of Big Data. Using examples of research projects from the INSITE center (www.insiteua.org), we will examine how value can be extracted from Big Data by employing different analytical techniques. In particular, the focus will be on large scale network analysis and visualization techniques, to understand relationships among the various types of data and develop prediction models. Our examples will span a number of areas including, healthcare, online news propagation, and education. The talk will highlight promising directions for research using Big Data analytics.