In This Issue:
Not Just Technical: Diverse Skills Fuel Data Science Success
Technical skills, such as programming computers and building predictive models, are often viewed as the main ingredient for success as a data scientist. But, far more is required to get the best results from a data driven approach. In order to get the most benefit out of data, the human element and insights that go beyond the technical are needed.
Over the last few years, the value of going beyond the technical has been getting greater recognition. Previously, with the rise of Big Data and machine learning, where computers can train to generate their own algorithms without human intervention, data scientists were often perceived as concentrating on the technical. But, the skills of a data scientist are not a one size fits all emphasis on writing computer code and developing algorithms. Insights based upon a broader understanding of what is being analyzed are also useful. And, those insights can unleash tremendous value from the data.
Many years ago, when I was that era’s version of what is today called a data scientist, I saw the limitations of algorithms that did not fully reflect a thorough understanding of the business dynamics. The value of going beyond the numbers is something I learned back then, which was a time when my expertise was heavily quantitative, and I was striving to broaden my business background beyond the strictly technical. Since then, I have devoted a good part of my life to understanding the patterns that drive successful business strategy. As a result of all this, I can personally relate to why a broader understanding beyond the technical can add tremendous value to data science.
And, it’s something I’ve written about previously in my October 31, 2012 blog post “Success with Big Data Requires More than High Tech Skills”, in my September 2013 newsletter article “Valuable Insights and Big Data Can Drive Success”, and in my December 31, 2015 blog post “More Support for the Shift toward the Softer Side of Big Data”. In my writing, I pointed out that not all data scientists must have computer programming skills, although many do. I pointed out that skills in understanding what the data really means are also valuable for data scientists. In other words, my previous writing explains that not all data scientists have the same skills.
More recently, the role of diverse backgrounds in data science was the subject of the March 13, 2017 Wall Street Journal article “How Do You Make a Data Scientist? Not Easily” by Deborah Gage. The article points out that “despite its name, data science isn’t just about being skilled with numbers.” But, the Wall Street Journal article actually adds an interesting twist to this. According to the article, “Every Friday, team members explain their projects and give one another feedback. Given all the skills a data scientist needs, Mr Hillion says, “you can’t have all that in one person.” Mr. Hillion is co-founder of Alpine Data, a “software startup that helps companies analyze their data to make predictions about their businesses.”
As I see it, Mr Hillion’s approach of having team members offer each other feedback has merit because a major challenge in data science is to bridge the gap between those whose background is primarily technical and those who are non-technical people with a greater understanding of the business. Bringing in diverse perspectives and adding a layer of human touch to the technical aspects can improve results and help to avoid situations where the algorithm goes troublingly wrong.
This human element can come from people who have subject matter expertise, such as the industry experience that Mr. Hillion tapped by assigning a data scientist who is a former derivatives trader to mine the data to find ways to prevent financial fraud. Or, the human element can come through people who are highly skilled at making sense out of the data, are very good at weighing why one dataset may be better than another, and can help make sure that the right questions are asked about the data. Essentially, various approaches to human intervention can work, since there is no one way to add this human element.
But, the human element is a crucial component of data science. And, the human element will remain important unless technology advances to a point where human intervention is somehow no longer needed. Despite advances in machine learning, however, that point is not yet here, nor is it likely near.
In some ways, the need for data scientists to transcend the technical parallels what has been the case for quite some time in disciplines, such as market research, for example, that relied upon data long before the advent of Big Data. Even back when data was collected with less sophisticated technology, there was value in being able to bring insights, rather than merely reporting the numbers. And, now that data is more voluminous, technology is more advanced, and the term data scientist has been coined, bringing insights continues to have great value.
So, in conclusion, it is important for data science to go beyond the technical. A variety of perspectives can contribute to greater understanding of what the data truly reveals. Data science is not just limited to mathematical and computer skills. Getting the most from the data calls for going far beyond that.
La Grange Park, IL