KDnuggets Home :: News » 2012 » Dec » Publications » KDnuggets Exclusive: Talent Analytics on Data Scientist traits and characteristics
Latest News


KDnuggets Exclusive: Talent Analytics on Data Scientist traits and characteristics

          
< Previous post Next post >

I spoke to Talent Analytics about the results of their recent study of Analytics Professionals and discuss some interesting results about characteristics of data scientists and how they work.

I have attended Dec 4, 2012 webcast which summarized 2012 Analytics Professionals Study The 2012 Analytics Professionals Research Study by Talent Analytics and IIA (International Institute of Analytics).

The study presented a number of interesting findings about analytics professionals, including data scientists, data miners, business analysts, and their managers.

I spoke to Greta Roberts and Jack Phillips, and the answers were provided by them and other Talent Analytics people.

Greta RobertsGreta Roberts is CEO of Talent Analytics, and a thought leader around on the inclusion of people in business optimization efforts. Her 20 plus years working with world-class innovators like Netscape, Lotus Development and Cisco Systems gives her critical experience successfully launching innovative software technologies into uncharted markets.

Jack Phillips Jack Phillips, is co-founder & CEO of IIA, which is his sixth successful start-up over the past 20 years. Jack specializes in spotting emerging job functions, and building successful information publishing and research firms to help those professionals make better decisions. Prior to joining IIA, Jack held operating and founding roles at INFONXX (now kgb USA), ISI Emerging Markets (now Euromoney, PLC), and CCBN (now Thomson/Reuters).

GPS: The "Data Scientist" job is currently a very hot job - recently described as the sexiest job of the 21st century by Tom Davenport and DJ Patil.
How much of "Data scientist" role is doing new functions and how much is"rebranding" of old titles like Statistician, Analyst, Data Miner ?

We believe much of what Data Scientists are asked to do today - are tasks they've been asked to do for years. However, today there are additional parts of this role because of the introduction of new technologies and disciplines like Machine Learning and Big Data.

Machine Learning and Big Data require more programming which is what we see many of the younger respondents in our Research spend a good portion of their time doing.

GPS: In your Research we saw a surprisingly small penetration of Hadoop - only 1%. How big among today's analytics professionals is the role/impact of Big Data (e.g. Hadoop, Cloud) vs. more traditional "ground" analytics ?

While there were no questions specifically designed to isolate "Big Data" users from the rest, we did pick up some of those people based on the tools they use.

Out of 302 respondents, 3 people selected Hadoop as a tool they use Other tools showed up on the "selected tools list" that imply big data. This tools list included: Netezza, Teradata, Splunk, mongoDB and Adobe Omniture.

We find the low level of selection interesting in comparison to the hype surrounding Hadoop. Our opinion is that many organizations are able to meet many of their analytics requirements using descriptive statistics like R or SAS.

GPS: How representative do you feel are your results, given a sample of about 300 data scientists?

There could be some biases but it's hard (or impossible) to know since we don't know the entire population of analytics professionals. As far as we know, there isn't another analytics professional dataset to compare our data to at this point.

One factor we noticed was that there seemed to be a decided lack of analytics bioscience people in our data. This stuck out for us given that we are headquartered in Cambridge MA where there are a lot of Bioscience firms.

We don't know why. Perhaps they go by a different title. But this is one area that stuck out for us.

Another area of "bias" was that our largest focus was a business focus vs. an academic or gov't focus.

GPS: One interesting finding is that 80% of analytics professionals work in small groups of 11 people or less. What are the implications?

We were somewhat surprised to see such small groups in analytics professionals even in very large organizations.

One possible reason for small group size might be that analytics professionals seem to be decentralized; embedded in departments vs. centralized in a larger analytics organization. So even in larger organizations with larger numbers of analytics pros - they tend to be distributed, in small groups, widely across the organization.

Implications could include:

  • The psychometric profile of the Analytics Pros we Studied are actually perfect for working in smaller teams.
  • They do like to collaborate and connect with others - but in smaller teams
  • They are not very "Politically motivated" meaning they could toil in obscurity unless they work for a leader that can promote their findings and advocate for their analytics work and findings.
  • Could this data be suggesting that businesses are generally under-resourced? Can businesses be analytics competitors with teams as small as 11 or less?
GPS: You identified 4 functional clusters: Data Prep, Programmer, Manager, and Generalist. What are some of the surprising findings regarding these clusters?

Talent Analytics 4 functional clusters

We had assumed that we'd see more "Generalists" in smaller organizations where there were requirements to do everything in the analytics pipeline. We expected specialists to emerge in larger organizations. This isn't what we saw.

Even in larger organizations Generalists were prevalent. We believe this ties in with our findings that groups of analytics professionals tend to be smaller - where the requirement persists to do much in the analytics pipeline.

GPS: Please tell us about psychometric characteristics of analytics professionals

Let us be clear - we found that there is more than 1 "type" of Analytics Professionals.

But, as a start let us talk about the Median as a way to describe the AP in our Study.

The Analytics Professionals in our 2012 Study are very creative and curious. They have an insatiable thirst to learn; wanting to look everything up on the Internet immediately because they simply "have to know". You'll find them at Meetup Groups on the weekends because they are so compelled to learn more. It's no wonder they're easily bored when they do nothing but data cleansing.

They won't be satisfied with their first solution and will continue searching; using their powers of imagination to continue exploring till they've exhausted all possible angles.

They are also a group of realists; calm and objective in a high-pressure situation. When something doesn't work, they'll assume there is a rational explanation and jump in to find and solve it.

You won't tend to find them sharing much their personal lives. Most conversations will be primarily about things that are less personal. They are a more reserved group, may take awhile to warm up and will be most comfortable with smaller groups; perhaps their inner circle.

GPS: What does Talent Analytics offer?

Talent Analytics has a mental model for understanding human traits as they relate to business performance. Our SaaS software measures these traits and reports them as 11 numbers that can be used in a variety of operational and analytical scenarios.

GPS: How does it help companies find their analytics talent?

These metrics allow analysts to create benchmarks of human traits that drive business performance. These benchmarks are used to optimize Operations and Hiring.

GPS: Here are The 2012 Analytics Professionals Study results (registration required).


< Previous post Next post >

KDnuggets Home :: News » 2012 » Dec » Publications » KDnuggets Exclusive: Talent Analytics on Data Scientist traits and characteristics

Copyright © 2012 KDnuggets.  | SUBSCRIBE to KDnuggets News email  | Tweet Twitter | facebook Facebook | RSS RSS | About KDnuggets