GovLoop - Knowledge Network for Government

The Next Generation of Analytics: Insights from Dr. Michael Evanchik

“Data analytics should be a core competency taught to all college students.” –Adrian Gardner, Chief Information Officer, Federal Emergency Management Agency

 

Only 15% of government agencies rely on analytics to meet agency goals, even though nearly 85% of agencies acknowledge analytics can bring increased efficiency and productivity to their agencies, according to GovLoop’s most recent survey.

 

One of the obstacles standing in the way of harnessing the power of analytics, cited by Gardner and our survey respondents, is the lack of employees trained in analytics. Gardner calls this knowledge gap “data illiteracy.” Even though industry sees analytics as a critical field, progress has yet to be made on incorporating analytics into educational programs.

 

To discuss that knowledge gap, I spoke with Michael Evanchik, PhD, the Associate Dean of University of Maryland, The Graduate School at University College (UMUC). UMUC has recently launched a Master’s Program in Data Analytics to meet this critical need. “The idea came from a conversation with UMUC’s own analytics staff. They brought up the fact that there aren’t enough programs out there to train people,” said Evanchik.

 

How can we train talented students and current employees to produce a generation trained in analytics? Evanchik lists 5 ways.

 

Understand your agency’s mission before diving in to analytics

“We see the role of an analyst as someone who understand the operational environment of their agency, the problems and issues their managers are facing, so they can help those managers get the kind of information in the proper format that they can then use to make better decisions,” said Evanchik. Analytics are a means to an end. Top data analysts must understand the goal of their agency in order to provide the right kind of data for decision makers. Once analysts understand what executives need from data, data experts can make a larger impact with the results of their analyses.

 

Train analysts in communication and business management, not just technology

The ability to turn data into actionable information is a skill that involves communication, business and technology processing skills. Most often, these skills are taught separately, making it difficult for analysts to combine these competencies together to apply toward analytics. In the UMUC program, students are put in virtual lab environments to test these skills. “They need business competencies, technology competencies, and they need computational competencies,” Evanchik said of data analysts. All of these components are addressed in the virtual lab. Later in the program, students actually work on a real project in an agency to apply their skills to a project. Evanchik says real projects are especially critical because data analysts must “understand the technology of data management. They have to be able to speak business. They have to be able to speak operations. They have to be able to communicate all of that. They, also, of course, have to do the analysis part.” Make sure any training in data analytics involves more than software proficiency.

 

Know the level of training you need

PhD programs in computer science usually take five or six years. PhD programs that focus on analytics train students to write the algorithms that form the backbone of the theories of data processing. While Evanchik himself has a PhD, he says that most agencies do not need someone with such high-level, theoretical skills. “Our focus is having our students use the algorithms that are already built into industry standard software,” Evanchik states. For the vast majority of projects, analytics code that has already been written and included in software is more than enough to accomplish key missions. Employers should focus on training that emphasizing leveraging existing tools, instead of long costly programs to invent new ones.  

 

Give analytics a personality

Just as analytics can give agencies insight into their customers, data analysts should look into their own personality strengths and weaknesses in order to tailor their responsibilities. Evanchik recommends anyone going into analytics take a personality assessment to see if their temperament is right for the role. “We see the role of an analyst as someone who understand the operational environment of their agency, the problems and issues their managers are facing, so they can help those managers get the kind of information in the proper format that they can then use to make better decisions.” That skill set takes a certain type of personality. The UMUC program uses the Birkman Method.

 

Analytics are a matter of “when,” not “if”

Training students and employees in analytics is a must for every kind of agency. As Evanchik pointed out, “There is something to do for any person with a strong analytics background in every organization.” Analytics most likely has a role in every agency’s mission.

 

 

What do you believe should be the priorities in training data analysts? What has worked with your agency? 

Views: 817

Tags: analytics, career, human resources

Comment

You need to be a member of GovLoop - Knowledge Network for Government to add comments!

Join GovLoop - Knowledge Network for Government

Comment by Mark Hammer on October 17, 2013 at 9:50am

There was a provocative article in Fast Company a few years ago called "Why We Hate HR".  And one of the themes running through it was that HR was often unable to sit at the grownups table largely because it was weak on analytics, and unable to offer up empirically-supported strategic ideas.

I think one of the important aspects of analytics is that it should be prospective, and not just retrospective.  That is, not just reactive to the information/data at hand, but linked in enough with what lies around the bend for the organization to tailor info-gathering to suit future analytic needs.  I know when I am involved in designing corporate surveys, I try to anticipate what could be done with the information gathered, to address future needs.  That consists not only of the content domains addressed (i.e., we'll need to know about X in 5 years), but also anticipating what can be done quantitatively with the data, knowing what sorts of statistical strategies can be applied if one gathers the information this way rather than that way.

And for that function, you can't just stick in a statistician or programmer.  You need a full-on analyst who is linked in with management and has a deeper understanding of corporate needs.  You can certainly grow those within the organization, given time and the right infrastructure, but Evanchik is onto something in proposing formal training programs to develop such folk.  I would imagine many academic programs in program evaluation could be easily tweaked in such a fashion, if they aren't already oriented that way.

Comment by Bobby Caudill on October 17, 2013 at 9:24am

Kathryn, 

I believe Dr. Evanchik's advise is refreshingly pragmatic and devoid of the hype found in so many articles and opinions. It's nice to see UMUC is taking a real-world approach to Data Analytics.

I recently jotted down a few of my own thoughts with regards to the expectations being set by the "Internet" for data science and data scientists. Perhaps your readers might benefit from the analogy used in my post.

http://bit.ly/197AWLJ

Comment by Henry Brown on October 16, 2013 at 2:57pm

New Site which could have some interest/value to some:

http://analystone.com/

About Analyst One

Analyst One is a site created by and for analysts

Our Mission: Serve analysts working hard problems.

Our Goal: Provide content designed to help analysts improve their ability to understand.
Topics Covered by Analyst One include:

    Analytics: the use of advanced tools and methodologies to extract meaning from data.
    Sensemaking: advanced concepts to help organizations create knowledge in systemic ways.
    Megatrends in the Analytical Community: The major forces in the analytical community that will impact us all.
    Technologies: Tools that can improve analytical outcomes, especially user-focused applications.
    Architectural constructs: For enterprise support to the analyst.
    Analytical challenges and use cases: The hard problems that need more community focused thought, and repeatable solutions that should be broadly shared.

© 2014   Created by GovLoop.

Badges  |  Report an Issue  |  Terms of Service