February 2, 2015 at 9:05 pm
Comments posted to this topic are about the item It's OK to be an Outfitter, not a Prospector
February 3, 2015 at 2:49 am
I think that the author is being a little hard on himself just because he didn't strike gold on that occasion, but is correct in pointing out that mining is just that - mining, and so there simply may not be any mineral wealth in that particular spot.
It may well be that moving to another claim is what is required, or a spell in the local saloon, chatting to the other prospectors. To those who have the time and inclination in pursuit of digital treasure, I say 'Keep digging'.
February 3, 2015 at 3:04 am
its 10am London Time
I have just arrived at the office
I am currently laughing my head off.
Great analogy
February 3, 2015 at 5:44 am
stephen.kemp (2/3/2015)
I think that the author is being a little hard on himself just because he didn't strike gold on that occasion, but is correct in pointing out that mining is just that - mining, and so there simply may not be any mineral wealth in that particular spot.
I think there is a bigger point here: Be aware of your limitations. Data Science is somewhat like building data warehouses: There are a lot of ways to go wrong. Data Science and Predictive Analytics are demanding disciplines. The author's point (which mirrors my own) is that as data professionals we need to have sufficient grasp of them to support users who possess the requisite knowledge and ability.
For the last year or so I've developed a solution to capture certain labor data, mostly so we can more accurately bill our clients. I borrowed a page from data warehouse design and concentrated on capturing the data at an atomic level so it can be rolled up and analysed any way end users need it. So now when our people who are versed in analytics want to slice and dice it, we get together and talk things over so I can work up a dataset for them.
I also have an extensive linked Excel workbook that I peruse every day and if I see anything interesting or odd I inform the appropriate users and let them take it from there.
____________
Just my $0.02 from over here in the cheap seats of the peanut gallery - please adjust for inflation and/or your local currency.
February 3, 2015 at 6:08 am
My point was that you should not regard attempts to journey further afield into other disciplines that don't produce an immediate result as a failure. The author got quite a lot from this exercise, not least of which, I suspect, is the ability to communicate effectively with data scientists and domain experts on the capabilities and prerequisites of data mining. And also learnt that data mining was not for him. My currency used to include halfpennies, and I would not be an IT professional today if I had not strayed from my earlier career as a metal basher. So I am sticking by my original assertions.
'There's gold in them thar hills' may or may not be true, but unless you make the journey there, you'll never know 🙂
February 3, 2015 at 6:26 am
Know thyself.
February 3, 2015 at 9:53 am
Thanks, and you are absolutely correct. In this one limited instance I had a complete miss. However I did learn tons about the tools, algorithms, and skills. From that I have been more able to communicate with other data professionals and "data scientists".
And I probably need to clarify, I do occasionally look at http://www.kaggle.com/ and play with different tools like R, Weka, and good ole Analysis Services. When it comes to associative, classification, and other algorithms I have skills.
What I fear are newly minted DBAs getting down on themselves because they don't have a PhD in statistics.
February 4, 2015 at 1:57 pm
Thanks for the article Ben. I'm sorry that you did not strike gold on your search. Yet you may have turned up nuggets even so.
You got me thinking of things in terms of the mining analogy. If we look at mining techniques in terms of waste and side effects we can see how some of the things we do to a database are just as drastic. Ever heard the old argument "you can't mine with a fire hose"? Well it's been done. It's called hydraulic mining. High pressure hoses were aimed at a piece of ground and washed the landscape into a run off. The fluid was filtered and the metal was removed. Everything else went down the drain.
One parallel is the tendency to stick all filtering in the WHERE clause. I have seen too many times where looking at a slow running query and examining the plans only to see fat lines on the right that were filtered down moving to the left. How can we get the metal without digging up the whole mountain? In these cases it was moving the predicates to the JOIN clause. Later we made even more improvements using things like CTEs to pre-filter or pre-aggregate.
So in this case the mining analogy holds. How many times do we dig up more than we have to only to waste memory or space in TempDB?
ATBCharles Kincaid
February 4, 2015 at 2:53 pm
Prospectors come and go, but the outfitter seemed to always make a living. Pretty much guaranteed reward, lower risk, no huge pay out.
February 13, 2015 at 12:02 am
The premise that your exercise in analytics was a failure only holds true if there was something to find. If there wasn't then failure to find something of value is not really a failure.
Gaz
-- Stop your grinnin' and drop your linen...they're everywhere!!!
Viewing 10 posts - 1 through 9 (of 9 total)
You must be logged in to reply to this topic. Login to reply