discovering hidden value through datamining

First off, everything I will blog about will be found here; http://www.thearling.com/text/dmwhite/dmwhite.htm http://www.thearling.com/text/dmtechniques/dmtechniques.htm

Anyway what, this blog is about is HOW we discover hidden value through datamining. Essentially, this refers to how datamining "scours databases for hidden patterns, finding predictive information that experts may miss because it lies outside their expectations". In the last blog everyone (as in the datamin class) talked about what datamin is. In my last post I talked about how datamining could be used by businesses as a tool to make them more efficient. This is what those hidden patterns and other predictive information do. for example, by being very shrewd, a person can predict demand of their product over a series of time and can produce in such a way that will maximize his profit by making the most of that demand. This is where the datamining makes money, by taking in more and more data, a person can then identify trends in the data collected. Seasonal things for example is a trend, for example like winter clothing being "in" during winter season (obviously). But, that is a simple exam people have more trends than that and are a lot less obvious such as fashion trends. In an ideal world, if you had enough data you could make a plan to maximize efficiency and profits in your business. Lets say your Mcdonald's, you could calculate when would be the best time (let's say season) to bring out your promotional foods, like twister fries. Specifically, you could decipher when people will most crave the product using past data to determine trends over several lengths of time.

So far, I've only mentioned a very single minded way of using the data for certain trends such as seasons. The application of datamining covers much more than that however. You could see trends and other predictive data in much larger scales. Such as seeing oh how men will react to your product as opposed to women, how african americans like it compared to asians and many other things. You could then choose to tailor your product to be most attractive to a certain race, sex, age, etc. Having this information will make it that much easier to plan and maximize whatever it is you do for business. The only thing that will be different among business is what their using this tool for. A business that will make some product will probably look at it's appeal to different groups. A business centered on service will look for where they are most wanted, for example, a massage company might want to relocate to a more industrial area where more people will be wanting their services after a hard day's work. An entrepreunership business will be looking for demand and be looking to supply to that place or groups of people to maximize their sales. The entertainment business will be looking for the fans who will be most receptive of whatever brand of entertainment they bring. The sports industry will be looking for mostly the same thing, for example, the NBA only sets up their teams to states in the U.S. where there are the most basketball fans, they have even stated that they might set up internationally if their sport gets more popular there.

This data will deal with mostly statistics. Simply put, a large count of people liking this and that. Arranged in such a way that there are several inferential data such as race, age, gender, sex (as mentioned earlier). Even without a complex program or mechanism that will simply tell you "oh this is the best choice for you". You could simply sort through the data to find which of the groups simply have the largest number of people which like or dislike whatever your selling. Like I said in my previous post; "knowledge is power, guard it well." All this data would be irrelevant if it could be found anywhere (shared on the internet) so in essence, this tool would only be helpful if only you and only you (and your co-workers) knew about it. This is probably one of the reasons there are laws such as patenting.



These days, there are subjects like gametheory, production control, operations management, etc. that deal with the application of the large amounts of data about people and other things. Datamining can be though of as the gathering tool and these subjects I have just mentioned as the application tool. I mention this to strengthen my next point, without datamining those application tools/subjects would have no value whatsoever.



I will end my blog here, with a little summary. How you discover hidden value is the next step in datamining, most of us in the datamin class probably talked about how datamining is a great tool to gathering large amounts of data. The next step is what all that gathering is for, to make inferences from statistical data to make predictions and to see trends. This can then be used as a great application to your business, assuming other people don't have you and your business have you in their own data (by knowing what data you have then taking advantage of that).

Again, "Knowledge is power, guard it well". :P