Tek-Tips is the largest IT community on the Internet today!
Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!
-
Congratulations derfloh on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!
Data mining being applied to... ? 3
- Thread starter Predictor
- Start date
-
1
- #2
I spent some time analysing what the difference is between San Miguel, Peroni, and Heineken beer drinkers in Italy. I think the answer was: not a lot!
Much of my work has been on consumer data and mainly the results of the analyses was a greater understanding of how the data was captured and manipulated before I got it. This is a useful outcome, but not quite what I was expecting. It was quite hard to produce any 'real' answers because as the granularity of the data became finer, the skew effects of capture techniques starting showing up. My home ----> visit me for Java and Data Warehousing resources
Much of my work has been on consumer data and mainly the results of the analyses was a greater understanding of how the data was captured and manipulated before I got it. This is a useful outcome, but not quite what I was expecting. It was quite hard to produce any 'real' answers because as the granularity of the data became finer, the skew effects of capture techniques starting showing up. My home ----> visit me for Java and Data Warehousing resources
- Thread starter
- #3
I was doing OLAP with a Business Objects front-end to a star-schema design Informix database. Then I started using Angoss KnowledgeSeeker to mine it. I wasn't too impressed with the BObj Miner at the time.
I can't say that we got into the data mining side that heavily as it was the beginning of a trial project. My home ----> visit me for Java and Data Warehousing resources
I can't say that we got into the data mining side that heavily as it was the beginning of a trial project. My home ----> visit me for Java and Data Warehousing resources
- Thread starter
- #5
I attended a training course and used it a bit. I enjoyed using it and found it flexible to use. Lots of people prefer SPSS or SAS, it depends what you are used to....and how much statistical knowledge you have. My home ----> visit me for Java and Data Warehousing resources
- Thread starter
- #7
Susan Shepard (over on the AI Discussion Forum at DDJ) responded:
"A company I am doing some work with now is playing around with customer defection prediction. I'll ask how much I may share about this at present and report."
That's interesting. I conducted an informal survey of data mining applications by talking with data mining tool and services vendors and examining job listings in the data mining field and concluded that roughly half of the work being done in data mining is applied to marketing or other customer behavior analysis, and about half of it is being done for the money industry (banking, finance and insurance).
"A company I am doing some work with now is playing around with customer defection prediction. I'll ask how much I may share about this at present and report."
That's interesting. I conducted an informal survey of data mining applications by talking with data mining tool and services vendors and examining job listings in the data mining field and concluded that roughly half of the work being done in data mining is applied to marketing or other customer behavior analysis, and about half of it is being done for the money industry (banking, finance and insurance).
-
1
- #8
Hi,
You might want to look at
and select what industry you want to know about.
Some of them have actual teradata customer stories about how they are using teradata and teradata warehouse miner (teraminer) product.
For example under
Industry
Energy
you will find....
Info-Lode: The Benefits and Success Factors of Data Mining in the Energy Industry.
It is a PDF document you can examine and/or download.
(Most of them have copyright provisions so be sure to honor the authors intentions if you intend on redistributing them.)
They may be more but I didn't go through all 7 industries teradata Supports.
hope this helps.
You might want to look at
and select what industry you want to know about.
Some of them have actual teradata customer stories about how they are using teradata and teradata warehouse miner (teraminer) product.
For example under
Industry
Energy
you will find....
Info-Lode: The Benefits and Success Factors of Data Mining in the Energy Industry.
It is a PDF document you can examine and/or download.
(Most of them have copyright provisions so be sure to honor the authors intentions if you intend on redistributing them.)
They may be more but I didn't go through all 7 industries teradata Supports.
hope this helps.
- Thread starter
- #9
What are the most novel data mining applications you've heard of? I recently read that during Clinton's run for re-election, his staff used a neural network to identify swing voters. Supposedly, they turned out to be voters in 2-parent households who bowled (?).
I've also read about using neural networks to grade the poppability of popcorn kernels (from machine vision images).
Anyone else?
I've also read about using neural networks to grade the poppability of popcorn kernels (from machine vision images).
Anyone else?
I used to run a marketing database for a cruise firm. To show them that they shouldn't always act upon what their data 'tells them' I produced a study on 'star sign' and 'propensity to cruise'. I discovered that water signs (Pisces, Aquarius, etc.) were more likely to want to sail.....but only very slightly...so I gave them a non-zerod y-axis graph to overemphasise the differences. My home ----> visit me for Java and Data Warehousing resources
-
1
- #11
I spent most of my time anylizing financial data and employees performance, it is very chalenging specially on the employee data where there is not a lot to compare to, and the DW is just now getting enogh data to represent the poplations. AL Almeida
NT/DB Admin
"May all those that come behind us, find us faithfull"
NT/DB Admin
"May all those that come behind us, find us faithfull"
I worked for a medium size manufacturing firm and found some interesting data on smoking. If you were a smoker, you were required to clockout into "Personal Time". "Personal Time" was considered a non pay activity. When I was running data on "Personal Time", I came across some interesting statistics. Employees are allowed to take smoke breaks when they feels like it. So it wasn't like everyone took them at the same time. I could forcast production volumes based based on employee smoke breaks.
Anyway, one woman asked me to run a report to find out how much income she was loosing due to smoking. Two days after I ran the report, she quit. Work was getting in the way of her smoking. She was loosing 10K a year on smoke breaks.
Anyway, one woman asked me to run a report to find out how much income she was loosing due to smoking. Two days after I ran the report, she quit. Work was getting in the way of her smoking. She was loosing 10K a year on smoke breaks.
Well, after writing and deleting a reply to this thread several times, I've decided that all I can say is...
"We're using a very powerful data mining tool to predict customer behavior related to a specific business problem with GREAT success."
Yes, that's so vague that it's almost not an answer. I'd just LOVE to go on and on about specifics of what we're doing, because it is very cool, but it is considered a competetive advantage and going into detail would possibly be mitigating that advantage.
"We're using a very powerful data mining tool to predict customer behavior related to a specific business problem with GREAT success."
Yes, that's so vague that it's almost not an answer. I'd just LOVE to go on and on about specifics of what we're doing, because it is very cool, but it is considered a competetive advantage and going into detail would possibly be mitigating that advantage.
I wouldn't say that we are using data mining techniques yet, but have begun using regression analysis to identify causes for defective products. Process data is queried by part number, production unit, shift, you name it and then compared with the scrap for the same time frames. If I've learned anything, it's that regression analysis isn't enough. The data is stored in multiple databases and pulled into an ms access back end. From there vb.net does the rest. Does anyone have a suggestion for my next step. The users of the system like what they see, but the results aren't there yet. I keep wondering how inaccurate some of the data is and plan to do some design of experiments to better understand process variablity and measurement accuracy. I really need a system that can handle data that may not be perfectly accurate. Any advice would be well received. Thanks in advance, Fred
- Status
Similar threads
- Locked
- Question
- Locked
- Question
- Locked
- Question