kitchen table math, the sequel: 'you can't hide from math and statistics and be a good marketer'

Thursday, September 29, 2011

'you can't hide from math and statistics and be a good marketer'

In the growing field of data analytics, finding "qualified candidates has proven difficult" for employers.

Is data analytics the new 'plastics'?
Posted by at

9 comments:

Bonnie said...

My university is starting a program, which will be housed in the math/computer science department (where it belongs, IMHO).

One of the biggest changes in my field (computer science)over the last 15 years or so is the degree to which statistical reasoning has taken over. We used to argue whether CS students should dump calculus in favor of more statistics. Now, I think it is imperative that they take more statistics.

September 29, 2011 at 6:06 AM
ChemProf said...

Oh, how I wish I could convince our math/CS department of this.

September 29, 2011 at 9:15 AM
Bonnie said...

My problem is that it has been so long since I took statistics that I really need a brushup, but I can't find the time. Maybe I should volunteer to teach the course - that is often the best way to relearn a topic!

September 29, 2011 at 9:44 AM
Grace said...

Why do you think this program should reside in the math/CS department?

IBM has teamed up with 200 schools to develop analytics curriculum, and the article makes it appear it's with business departments. I noticed that some business schools offer concentrations in IT management.

September 29, 2011 at 12:13 PM
Bonnie said...

Because this has been a hot field within CS for about 10 years and there is a lot of knowledge in the community. I think a lot of the fundamentals have been developed by CS people, and many CS programs already do courses in the field.

And, according to our wonderful expert Wikipedia "Analytics therefore bridges the disciplines of computer science, statistics, and mathematics" (actually, that quote comes from a 2002 article in Communications of the ACM, which has been publishing articles in the area for years). That sounds awfully like what we teach in a math/CS department, hmmm?

September 29, 2011 at 3:36 PM
Bonnie said...

Some of the schools that house data analytics in CS
Illinois Institute of Technology (Masters in CS with a specialization in Data Analytics)

Virginia Tech teaches a grad course in Data Analytics in the CS department

DePaul's program, which is in conjunction with IBM, is in their College of Computing and Digital Media.

Georgia Tech includes Data Analytics, Machine Learning, and Visualization as a core area in their School of Computational Science and Engineering.

I am sure there are more examples - I just grabbed those on a quick web search.

University of Vienna houses their Data Analytics and Computing research group in the Faculty of Computer Science

here is what Stony Brook is doing
http://www.cs.sunysb.edu/about/news/cdda.html

The field is fundamentally about computational methods on very large datasets. The research has come out of computer science, with a heavy dose of statistics.I think a data analytics program in a business school would be very applied and would have to focus on using tools and algorithms created by someone else, whereas a program in a CS and math department would focus on developing new methods and algorithms.

September 29, 2011 at 3:50 PM
Anonymous said...

"I think a data analytics program in a business school would be very applied and would have to focus on using tools and algorithms created by someone else, whereas a program in a CS and math department would focus on developing new methods and algorithms."

If the target is people who will go into marketing, I'm pretty sure that what the (a) want, and (b) can handle WILL be very applied and should *NOT* focus on developing new methods and algorithms. The marketing folks are not going to be interested in arguing the advantages of boosting versus support-vector-machines ... or even in discussing how/when to use them together.

-Mark Roulo

September 29, 2011 at 6:12 PM
Bonnie said...

You are right. Think of the difference between computer science and MIS, which is housed in business schools. In computer science, students learn how to create computer applications, and to develop computational methods. In MIS, the textbooks have pictures of computers: "This is a PC. Here is a mouse. Mice are an input device". In other words, they read about computer applications rather than creating them. MIS was always the refuge for students who flunked out of CS.

September 29, 2011 at 6:15 PM
Anonymous said...

UCSC has no business school, but has a strong machine learning group split between computer science and "TIM" (Technology and Information Management), which is part of the Jack Baskin School of Engineering. The TIM program is a bit more engineering-oriented than a business school, and some of the faculty are quite competent in fields like optimization and machine learning.

September 29, 2011 at 8:30 PM

Post a Comment

Subscribe to: Post Comments (Atom)