I’m not a big fan of links-only posts, but I’m still chuckling at the “Danger Zone!” quadrant from Drew Conway’s post The Data Science Venn Diagram on the Dataists Blog:
Venn diagram by Drew Conway.
I just realize I posted to a cross-post, which is one of the reason I don’t like all this linking about. The original blog entry is on Drew Conway’s blog, Zero Intelligence Agents.
November 25, 2010 at 12:58 pm |
I have been quite surprised at the effectiveness of Danger Zone efforts in NLP. Generally it starts with just hacking skills and by the time enough data has been dealt with one has the Substantive Expertise.
State-of-the-art is now to move the Danger Zone to Data Science by transforming all those heuristic hacks into features for a well behaved statistical model. Logistic regression therapy for your inner hacker.
Breck
June 26, 2012 at 9:35 am |
Data can be fabricated, manipulated and altered. Even scientific data can be altered. What would be the outcome of the science of fabricated data?
June 27, 2012 at 1:45 pm |
If you fabricate data, that’s called “fraud”, not “science”.
What typically happens in practice is a long string of publications followed by a messy downfall.
In the long run, science is about replicable results and fraudulent results are unlikely to be replicable.
June 26, 2012 at 9:38 am |
Here some example.
http://en.wikipedia.org/wiki/Scientific_misconduct
http://www.dailymail.co.uk/news/article-1360490/Climate-change-scientists-accused-altering-data-cleared.html
http://juneauempire.com/stories/100606/sta_20061006008.shtml
etc.etc.