Are we getting lazy trying to solve quant analytical problems?
I used to use maths, write simulations and use statistical programs. now what I do is throw almost an variable I think of into a data mining software and let it work it out. Then I just work Do you think that is true for most people now?
==Given the sense of urgency to solving business problems, companies may not be patient to allow someone to “manually” select variables and build/validate models. I would not say we we are lazy. The tool could be entirely wrong becasue of the way the data was created or it does not generate results that make sense to the business instincts.
However, you still need all the math and stats talent to decipher the results. Sometimes, the tool, in my case SAS Enterprise Miner, may recommend a certain set of variables, but the business may disagree. As you know there is a lot of art and science in this work, so the tools help automate certain tedious tasks like varaible selection and model development/comparisons, but past experience/education is required to go with the course of action that makes the most sense to the business.
I laugh when people think anyone can throw data into a tool and spit out a model. Like any analytics tool, it’s only as good as the data its fed and there is no “magic” button.
==I agree with everything except that it does not save time, it takes more time. In my case it is formulize. A typical example might be I spend ages collecting data. then set it up for a four hour run but discover that I have made a basic error, so the next day I fix it up and run it again. Then I fix it up again, then run it over night. Discover some extra stuff. Then run it again and so on. It only stops when I finally give up and say it is good enough. Being a perfectionist that is a long time.
==I agree – this work is very iterative and unstructured. My boss gave me some good advice: “Don’t destroy the good for the best.” At some point, we have to realize what we have developed is better than what we started with and deliver it. Otherwise, we will never finish.== have a Professor that often says “Computing power is cheap” I personally think that with the advancements in computers and statistical packages, part of the job of the statistician is to find that balance between brain power and computing power that solves the problem most efficiently.
FACEBOOK ACCOUNT and TWITTER. Don't worry as I don't post stupid cat videos or what I eat!