Data mining, analytics, predictive modeling, artificial intelligence, natural language processing, database warehousing
Sep 4, 2009 at 8:02 PM Thread Starter Post #1 of 7

Welly Wu

Headphoneus Supremus
Joined
May 16, 2003
Posts
5,165
Likes
12
So, this is quite the subject matter for Head-Fi, but it does pertain to our niche clique. There have been tremendous research, development, and actual live platforms that deal with these topics and although the costs are astronomical, it is entirely feasible that the next two to five years will lead to a trickle down effect to real world prices and options. Here is a scenario: a newbie is looking for a great pair of headphones for under $50 USD and he is interested in reading a mixture of professional reviews both inside and outside of Head-Fi, subjective ownership reviews ranked by the "seniority" of the said author, and technical specifications along with how others are utilizing this aforementioned magical pair of cans in their audio equipment racks and specific applications. These technologies could help the so called newbie learn more quickly and get useful information more rapidly with satisfying real world results: he is a happy newbie to Head-Fi while the "veterans" are not pestered by anonymous private messages on ancient reviews posted years ago. Some of these platforms do coincide neatly and could be ported into the Head-Fi databases, servers, etc. Do you think that this would be a worthwhile medium to long term project? Who is in this for the long haul in terms of making contributions? I am already in it and learning these fascinating technical matters while combining my experiences as a Head-Fi member to tailor it to the needs of our community. It will take me time to be able to talk with Jude Mansilla and the other administrators about it and some time to put it online in a series of alpha, beta, and production ready tests. Whaddya think?
 
Sep 4, 2009 at 9:38 PM Post #2 of 7

Kees

Headphoneus Supremus
Joined
Feb 18, 2006
Posts
4,619
Likes
24
First you need to define the requirements: what kind of information/reports do you want?
You'd need an information analysis and a data model to describe what you need.
All data is in free (text) format, so you'd probably need agents to find the specific data, interpret it and translate it to a standard format.
Then you'd need to make functional descriptions of all agents and report generators and the user interface(s).

Just some things that pop up when I read your post.

Seems like doable, but not easy. And definitely long term.
 
Sep 4, 2009 at 10:09 PM Post #3 of 7

DeusEx

Headphoneus Supremus
Joined
Sep 13, 2007
Posts
1,939
Likes
16
Location
California
I'm very interested in studying those particular fields. Computer Science will take us farther and farther within the realm of "smart processing", or the semantic web.
 
Sep 5, 2009 at 11:28 AM Post #6 of 7

Lazarus Short

Headphoneus Supremus
Joined
Oct 23, 2003
Posts
3,363
Likes
13
Large retailers, especially Wanker Mart, are already doing this kind of thing. WM has a HUGE data facility in McDonald County, Missouri which is actually rather hush-hush. The single building and its contents was so costly that it doubled the county's tax base, or so I was told. Here's a relevant piece:

Wal-Mart's data center remains mystery


The Joplin Globe

By Max McCoy


Globe Investigative Writer

JANE, Mo. - Call it Area 71.

Behind a fence topped with razor wire just off U.S. Highway 71 is a bunker of a building that Wal-Mart considers so secret that it won't even let the county assessor inside without a nondisclosure agreement.

The 125,000-square-foot building, tucked behind a new Wal-Mart Supercenter, is only a stone's throw from the Arkansas line and about 15 miles from corporate headquarters in Bentonville, Ark.

There is nothing about the building to give even a hint that Wal-Mart owns it.

Despite the glimpses through the fence of manicured grass and carefully placed trees, the overall impression is that this is a secure site that could withstand just about anything. Earth is packed against the sides. The green roof - meant, perhaps, to blend into the surrounding Ozarks hills - bristles with dish antennas. On one of the heavy steel gates at the guardhouse is a notice that visitors must use the intercom for assistance.

What the building houses is a mystery.

Speculation

Wal-Mart's ability to crunch numbers is a favorite of conspiracy theorists, and its data centers are the corporate counterpart to Area 51 at Groom Lake in the state of Nevada. According to one consumer activist, Katherine Albrecht, even the wildest conspiracy buff might be surprised at just how much Wal-Mart knows about its customers - and how much more it would like to know.

"We were contacted about two years ago by somebody who runs a security company that had been asked in a request for proposals for ways they could link video footage with customers paying for their purchases," Albrecht said. "Wal-Mart would actually be able to view photos and video of customers paying, say, for a pack of gum. At the time, it struck me as unbelievably outlandish because of the amount of data storage required."

But Wal-Mart, according to a 2004 New York Times article, had enough storage capacity to contain twice the amount of all the information available on the Internet. For the technically minded, the exact amount was for 460 terabytes of data. The prefix tera comes from the Greek word for monster, and a terabyte is a trillion bytes, the basic unit of computer storage.

Albrecht, founder of Consumers Against Supermarket Privacy Invasion and Numbering, said she never could confirm the contractor's story. That is not surprising, since Wal-Mart seldom comments on its data capabilities and operations.

A Globe request for information about the Jane data center was referred at Wal-Mart headquarters to Carrie Thum, a senior information officer and former lobbyist for the retailer.

"This is not something that we discuss publicly," Thum said. "We have no comment. And that's off the record."

Read more >>
 
Sep 6, 2009 at 10:03 PM Post #7 of 7

Welly Wu

Headphoneus Supremus
Joined
May 16, 2003
Posts
5,165
Likes
12
That's right: linked data. Exploring the relationships among data sets, information tags, relational databases and the ever critical search indices all the while making it simple, human, and natural for end customers is the future of these closely interrelated academic disciplines. This stuff is so complex and fascinating in of itself. However, this is what we we do on an almost intuitive level especially for women and language. We are really talking about balancing the differences in the ways in which men and women view, use, re-use, and explain data. That is why I am so damned interested in this stuff.
 

Users who are viewing this thread

Top