Can IBM Watson Analytics Replace the Data Scientist?

By   /  October 18, 2016  /  No Comments

pg_watson_101816At the core of IBM’s Watson Analytics is a powerful bull’s-eye tool; it employs a powerful user query as the basis of analysis. The Data Scientist or business user is free to play with the variables in any dataset and directly visualize how the changing user inputs impact the answers. The bull’s-eye also helps narrow down the user query to help get a much clearer answer.

This kind of focus can be especially useful in a large dataset, where the user is likely to get lost. IBM’s design philosophy behind Watson was to keep the average user sharply focused on the query. Watson Analytics combines the power of Cognitive Computing with the flexibility of Natural Language Processing (NLP) to answer business queries.

With such promising capabilities, Watson Analytics brings up the obvious question – can semi or fully automated Data Analytics replace the Data Scientist? The IBM Watson vs. Data Scientist debate has recently gained some ground, although most of the credible voices in the industry seem to feel that the hype is far from truth.

In IBM Watson – Will It Replace Data Scientists, readers will discover that a 2014 review of IBM’s Watson Analytics (Beta version) revealed that this service was slated to automate the Data Scientist’s job function. This robot was designed for ordinary business users who aspire for something beyond routine spreadsheet analysis. Although Watson can potentially role play the “Data Scientist” in an organization, does that mean this machine analyst can actually replace the human expert?

Some years ago, the Beta version was tested for customer retention in the telecom industry. During the test, The “Predict” and “Explain” functions, in absence of human intuition, spit out all the findings from a dataset in order of statistical significance. This indicated that the relative significance of a business insight was determined by statistical methods, rather than by the cognitive processes of the human brain. Future versions of Watson may include storytelling and forecasting, which will perhaps strengthen the evaluative power of this solution.

The Truth behind the Hype

In Analytics-as-a-Service: Will Watson Analytics Replace the Data Science Role?  the global Data Science community expressed serious concerns about Watson Analytics gradually replacing Data Science professionals in enterprises. Data Mania, during its market review of analytics-as-a-service, concluded that Watson Analytics cannot and will not replace the Data Scientist. The current limitations of this Analytics platform are:

  • The learning curve is too steep
  • Watson does not support data export to other applications
  • All the analytical smarts cannot possibly be codified in Watson

Such prohibitions make Watson unsuitable for the average business user. In the future, if IBM developers decide to reduce the current restrictions, then maybe ordinary business users will have the capability to export the clean data out of Watson to other friendly applications. On the flip side, the observed strengths of Watson Analytics are its data cleansing, data visualization, and data modeling capabilities, which any user can tap into without worrying about coding. Thus Watson wins on one score – it encourages non-programmers to dive into complex business data.

All said and done, Watson cannot replace the knowledge of mathematicians and statisticians that Data Scientists bring to the table. So, at best, Watson can only act as an accelerator to the Data Scientist. Business users who are also sound in mathematics and statistics will have an edge over their peers, as those two knowledge areas can greatly enhance the value derived from Watson Analytics.

So where does that leave the average business user? Even after retrieving clean and visible data from Watson, they would ultimately have to turn to the Data Scientist for the intricate coding required to extract value from the data. The technical skills of the Data Scientist can never be fully codified into a robot.

IBM Watson vs. Data Scientist: Watson as Data Science Accelerator

The inherent qualities of Watson make Data Analysis quick and easy for the Data Scientist:

  • Watson enables ordinary business users to process large datasets and visualize large amounts of data very fast
  • Watson delivers ready data models and even rates the results
  • Watson uncovers hidden insights from large datasets very quickly

Thus, the combined capabilities of Watson and the Data Scientist will actually attract and develop many business users into expert analysts over time.

Role of Watson in Analytics-as-a-Service

A 2014 study by Hypatia Research Group stated that IBM targeted both the Data Scientist and other business functions, which lack the analytics expertise of Data Science, for example in sales, finance, HR, or operations. To this end, the company designed Watson in a manner so that:

  • The natural language query method (NLP) supports non-technical users
  • The time spent on data preparation is greatly reduced
  • The time to generate data models is substantially reduced
  • The data refinery capability delivers analysis-ready data
  • Stock queries, predictions, and correlations hasten the data analysis process

Readers can also review the article 9 Ways Businesses Are Using IBM Watson  to understand how Watson is currently aiding the Data Science function in organizations.

Both Forbes and Data Science Central share common sentiments in Will We Soon No Longer Need Data Scientists, where the “unicorn” image of the Data Scientist has been demystified. Although this article claims that the best job in America for 2016  may be at risk of losing out to Artificial Intelligence, every enterprise knows and acknowledges that no robot can combine the comprehensive skills of programming, mathematics, statistics, Machine Learning, database management, and domain expertise in its fold.

So the fear of losing Data Scientists to Watson is not really well founded. In reality, what a solution like Watson Analytics can probably do is successfully tackle the skills shortage in Data Science by augmenting Advanced Analytics processes with semi or fully automated data tasks. Moreover, Cognitive Computing can help reduce the time taken to go to school and learn Data Science.
IBM’s Vice President for Watson Analytics and Business Intelligence, Marc Altshuller states:

“A traditional Data Scientist might receive training in R or SAS or whatever tool their school uses, but we found in the ‘citizen analyst’ area, they were often being given the wrong tools where they were required to guess the right answer, and then test their guess.”

IBM Watson vs. Data Scientist: The Selling Points of Watson Analytics

  • The NLP query language can aid the business user to frame right questions about the data
  • The powerful visualization capabilities can immensely enhance comprehension for both the Data Scientist and the ordinary business user
  • The semi-automated data query, data cleaning, and data modeling functions can substantially reduce the Analytics process and leave the Data Scientists to focus on the Analytics task itself
  • The assisted Data Analytics approach can convert hesitant users into Analytics experts, but with the help of Data Scientists

The article IBM Watson Analytics: The Data Scientist Accelerator claims that when IBM lunched Watson Analytics in 2014, the enterprise positioned this solution as a game changer in the fields of Data Science and Analytics. The aim of this solution was to deliver a one-stop solution for all basic Data Science processes from data clean-up to data visualization. So, in essence, Watson does not replace the Data Scientist; in fact, it frees the Data Scientist from tedious tasks to engage in more involved Analytics.

Watson is a true Machine Learning tool; the more it processes data, the more it learns about the data, and about the changes made to the data to eventually automate the data clean-up process.

The Perfect Match of Demand and Capability

The article Machine Learning as a Service explains so well that when the world’s data growth suddenly spiked exponentially, the global Data Management community was at a loss for some time.  Now, as database technologies and tools continue to evolve to combat the volume, velocity, and variety of data, Machine Learning experts can fully take advantage of technologies like Watson in building apps that continuously learn from the in-flow of user data, designing tools to detect fraud, and in delivering personalized shopping experience to online customers.

About the author

Paramita Ghosh has over two and a half decades of business writing experience, much of which has been writing for technology and business domains. She has written extensively for a broad range of industries, including but not limited to data management and data technologies. Paramita has also contributed to blended learning projects. She received her M.A. degree in English Literature in 1984 from Jadavpur University in India, and embarked on her career in the United States in 1989 after completing professional coursework. Having ghostwritten and authored hundreds of articles, blog posts, white papers, case studies, marketing content, and learning modules, Paramita has included authorship of one or two books on the business of business writing as part of her post-retirement projects. She thinks her professional strength is “lifelong learning.”

You might also like...

Thinking Inside the Box: How to Audit an AI

Read More →