A Quick Look at Data Science

An article in the Harvard Business Review called the Data Scientist the sexiest job of the 21st century. This never fails to put a grin on my face, because some of the work we have done here at iMovo has more than just shyly dipped a toe into the silvery waters of data science. Though I wish that this would, transitively, imply that I am the sexiest employee this century, I think I will settle for being able to have a lot of fun with data at the office.

So what exactly is this job that the Harvard Business Review got so hot under the collar for?

The work of a Data Scientist is best explained when contrasted against that of a Business Intelligence specialist. The majority of Business Intelligence work centres around constructing and maintaining data warehouses and its reporting modules. These warehouses are, in turn, filled with clean historical data that allows employees to analyse past activity. Historical data is rooted in certainty, and as the famous (mis)quotation goes, “therein lies the rub”.

The nature of predictive analysis and pattern identification is highly experimental. A data scientist would root around messy wads of data, searching for patterns and new insights. While Business Intelligence systems are built to answer the questions that business users have, Data Science explores the data and finds answers to questions that users never thought to ask.

An example of this is the discovery made by the data science team at Netflix, where it was found that customers who added a certain number of films to their wish list were more likely to become long-term customers. This kind of exploratory work requires analytical skills, mathematics, statistics, good programming skills and a good grasp of data management. It requires a more “scientific” approach to solving problems. “Data Scientific”, to be more specific.

business_intelligence

So what happens once a data scientist has had their way with data? Well, a number of things happen. A biblically verbose report could be written and sent around, which in turn could perhaps precipitate changes in the way a business operates. Taking the Netflix example once again, the service now encourages users to fill their wish-list and has a more comprehensive trial service.

The data scientist might even work with the Business Intelligence specialist to find ways of adding any new discoveries to the data warehouse. Some companies go as far as to build data products – that is, a function or a family of functions that are driven by data. The “People You May Know” feature in LinkedIn is an example of a data product.

So what about the skills required? After leafing about monster.co.uk, looking at job vacancies and noting the requirements, I collated the profiles of the ideal Data Scientist and Business Intelligence specialist. These are summarised in the Venn diagram below. Of course, this kind of analysis is not entirely accurate – after all, an advertised vacancy is not always an accurate representation of the underlying job.

Data Scientist

These skills corroborate everything that was written above. Data Science is exploratory, it delves into different areas and brings together a flurry of disciplines. Business Intelligence is more rigid – it focuses on curating data and serving it to the end-users in measured and calculated ways.

On that note, I should tell you all that I did not write this article to insinuate that one discipline is meant to replace the other. Even in the almighty age of Big Data, Business Intelligence is still relevant and supremely important. If anything, these disciplines are two sides of the same coin; unified in the call to empower an organisation with both hindsight and foresight.

No, I wrote this article to shamelessly tell you that the little Data Science I have done here at iMovo was bloody good fun. It’s that kind of fun that truly makes Data Science the “sexiest job of the 21st century”.