Of the 259 respondents, 23% were students. The degree breakdown is in the table below.
|
What is your highest education level? | |
| PhD in CS/Math/Stats (52) |  20.1% |
| PhD in another field (30) | 11.6% |
| MBA (15) | 5.8% |
| MD/JD or other advanced graduate degree (7) | 2.7% |
| MA/MS (119) | 45.9% |
| BA/BS (28) | 10.8% |
| less than a BA (8) | 3.1% |
Regional breakdown shows that Middle East and AU/NZ have the highest percentage of PhDs.
| Region (% voters) | % PhDs | % Students |
|---|---|---|
| US/Canada (52.5%) |  29.4% |
 20.6% |
| Europe (29.0%) | 36.0% |
18.4% |
| Asia (8.9%) | 30.4% |
13% |
| Latin America (4.6%) | 8.3% |
41.7% |
| Africa/Middle East (2.7%) | 57.1% |
28.6% |
| Australia/NZ (2.3%) | 50% |
16.7% |
Comments
Chris Barnes, Student?
I guess that in your terms, I am no longer a studen, as I finished my PhD in 1975. But, as I try to learn a major new area each year (language (e.g. Spanish or R), simulation or analytical package...), either formal (pay money) or informal, it hurts to classify myself as a non-student.
jgm
Yay! I remain the least formally educated data miner! I can't believe I'm the world's only "hobbiest" data miner though....
Wednesday, June 08, 2011, 1:56:01 PM
Gregory Piatetsky
@jgm what data mining-related education did you receive?
Wednesday, June 08, 2011, 2:14:30 PM
jgm
@Gregory
None, other than reading books like Witten and Frank's Data Mining, Koza's genetic programming volumes, Freeman and Skapura's neural network textbook, etc. and lots of work on my own data sets (sports statistics, weather data, etc.) with Rapidminer and other open source tools and programming with Pascal and Delphi (plan on learning R and python soon).
Data mining and machine learning offer the incredible possibilities of seeing the future, understanding the present and explaining the past and in my opinion shouldn't be confined solely to academia. Much like programming, it's a skill set great for those who like open-ended puzzles and could quite possibly thrive among "amateurs" if made more accessible. Heck, anyone who's handicapped a horse race has made use of past examples to forecast a future event and quite probably derived rules/algorithms from the past performance data. The websites mentioned here that offer prizes for novel algorithms or improved results are a definite step in the right direction to get more people interested, in my opinion. The available of free tools like R and RapidMiner and Weka is another great move to enable the casually interested to explore without major investment.
Saturday, June 11, 2011, 3:27:00 PM
Gregory Piatetsky
@jgm - seems that you have made good use of real-world education !
However, doing data-mining well is not easy - there are plenty of examples of social sciences who analyzes data badly - just search for "Decline effect" , so good education is important for separating valid results from random noise.
Saturday, June 11, 2011, 5:20:25 PM