Showing posts with label Big Data. Show all posts
Showing posts with label Big Data. Show all posts

Wednesday, May 25, 2016

Angry at BIGdata

Angry, I'm livid.! Bubba watson has just taken the lead at the Memorial tournament. I pick up my phone to see if he burdied the par 5 7th and I have to hit "Leaderboard" 5 or 6 times. Every time I log in, I have to again re-confirm that Bubba Watson is my favorite player. PGAtour.com feeds me scores from the Classic Tour, the Canadian Tour - they show me play by play on other players... but they will not give me what I want, what I ask for repeatedly.
I just sold an item on eBay for several thousand dollars - best ever. Do you think I can find my messages? On my Win7 machine, I've learned never to load the eBay home page - which perpetually loads images of items which I already bought on eBay. I have been on eBay since 1997, I know how to find what I'm looking for, I've bought hundreds of items. But I can no longer go there. I've even gone so far as to load the mobile site on my desktop - just to prevent it killing my computer.
Google, to its credit, has not junked up the screen... but on every tech search I try, I'm forced to restrict the time frame to one month or one year. Do you think all-powerful Google could add the time frame button to their completely empty page - given that I use it several times each day? Nope, they must not be paying attention.
What does this have to do with BIG Data. I'll tell you. Companies are consumed with trying to predict my behavior and especially what I might be interested in buying next. And they are neglecting and annoying me by ignoring what I try to shout to them (repeatedly).
Lets illustrate the challenge in predicting purchases. Think back to the last time you visited a big box store. Perhaps we talked outside and made a list of all the things you intended to buy on your visit. What are the odds that when you come out, you have exactly that list of items, no more, no less? In my case it is no better than 100 to 1. So if you, knowing all you know about you, cannot predict your own behavior, 5 minutes in advance. What chance does some computer have, merely looking at your lifetime tweets?
BIG Data is capturing the attention - and the result is anger - your customers are angry - they hate your new "Predictive" site. They hate that for weeks they see ads for items they already bought or weren't really interested in. But most of all, they hate that you took away their time and access so you could piddle with global data feeds. Give your customers what they want - but quit trying to out-guess them.

Sunday, January 27, 2013

BigData Interest Explosion!!





BIGdata Video

Because I frequent analytical groups, blogs and virtual cafe's, I have been inundated by the term “BIGdata”. And as the Greek poet Menander says: “Bad company corrupts good character.” my vision could be a bit distorted. I got so excited I even did two videos about it – my most watched. http://bit.ly/BIGdataJRM So when I decided to explore Google's new Search Trends I thought it would be fun to see the history of this term.

Chopping the scale off makes it look more impressive
This also gives you a nice view of how this new tool works. As you can see, BIGdata has gone from basically non-existent into the stratosphere! I was so excited to be part of this 'trending' (here's the trend on that). Well, as with most analysis, if it looks too spectacularly exciting, you might want to do some comparisons. I noticed that the scale was 0-100 and that seemed like a pretty low number... then I realized that this was a relative scale. So it can be millions and millions of searches – or not... hard to tell.
Chopping the scale off makes it look more impressive

There is an additional feature which allows you to compare a second, third or fourth (not sure the limit) search term. I thought that might give me some greater perspective on this exponential growth in BIGdata (of course it is BIG, look at how it is growing).







When we compare it with 'Database Marketing' we see that there has been a huge decline in what once was a trending phrase. I remember discovering that I could charge almost double if I talked about 'database' instead of the tired 'Direct Marketing'. Google doesn't offer trending before 2004 but this graph suggests that it may have peaked around that time. What the comparison graph shows is that even compared to a pretty 'niche' term like 'Database Marketing', BIGdata hasn't really gotten much use outside of our analytics cafe group.

As I mentioned in my video Data Science video, the first and most important character trait is curiosity... so of course I had to go deeper.

Even though most of my friends and readers will recognize the “Database Marketing” term, I decided to compare still a more general phrase to see what its scope might be.

You can see that the simple term, “Database” while declining, crushes the other two terms to such a large extent that both completely disappear. What this suggests to me is that many many more people are doing things with data and databases than are concerned about 'Database Marketing' and/or BIGdata. Which is a super great illustration of the problem.

We want to build data, work with data, pile up database but we really do not want to take much time letting marketing do much with it. Which is also why we find it far easier to make money with data when we do not work within the traditional database environment. And that 'out of the box' thinking is why we beat the best modelers in the world. So keep your attention on your work, your data and get some real data scientist to help you make sense of the back end.