Deprecated: Function set_magic_quotes_runtime() is deprecated in /home/mwexler/public_html/tp/textpattern/lib/txplib_db.php on line 14
The Net Takeaway: Page 33

OTHER PLACES OF INTEREST

Danny Flamberg's Blog
Danny has been marketing for a while, and his articles and work reflect great understanding of data driven marketing.

Eric Peterson the Demystifier
Eric gets metrics, analytics, interactive, and the real world. His advice is worth taking...

Geeking with Greg
Greg Linden created Amazon's recommendation system, so imagine what can write about...

Ned Batchelder's Blog
Ned just finds and writes interesting things. I don't know how he does it.

R at LoyaltyMatrix
Jim Porzak tells of his real-life use of R for marketing analysis.

 

HOW DID YOU GET HERE?

baidu.com
https:
google.ca
baidu.com
https:
https:
https:
https:
https:
https:

 

 

 

Bill Littlefield's talk in Needham · 04/10/2004 12:50 PM, Personal

Who is Bill Littlefield? You may know him from commentary on NPR’s Morning Edition. He also has his own syndicated NPR show called Only A Game. He talks of sports and the players, sometimes opinion pieces, sometimes journalism.

I find that when I hear him, I stop and listen. Sometimes on NPR, more often when I’m scanning the dial looking for Car Talk. Big deal, you say, lots of sports fans listen to him.

Only, I am not a sports fan. I get bored to tears trying to watch sports. Ignoring the ungodly costs of a ticket, I just find sports to be a waste of time. (Yes, I have the classic correlation of not being good at any of them, so of course I don’t enjoy watching them). I don’t care about the stats (one of the few datasets I find it easy to ignore), I don’t enjoy the celebrity, and I don’t appreciate the skills of the players.

But I find that when I start to listen to Mr. Littlefield, I understand what sports fans like. Mr. Littlefield brings out the adventure and the joy of competitng, not necessarily winning, and connects to the human aspects of these athletes who are often just caricatures in the media, or as flat as the card with their picture on it.

Last week, April 4, he gave a talk at the Needham, MA library. It was a full house (40 or so people, including kids).

When he reads, he sounds like he does on the radio: somewhat of a sing-song, a measured tempo reminiscent of a sportscaster who is no longer trying to prove that the sidelines media are as important as the game itself. When he’s just talking, he sounds less engaging, but still powerful. He’s a tall, thin fellow. He came very casually, in a sweater and pants. He knew folks in town, in part b/c he coaches a girl’s soccer team in the area. And folks came in from miles away, some an hour’s drive, to hear and see him.

He chatted about the first game of the Red Sox this year, the joy of seeing college ball and minor league ball, the high prices of tickets, the high pay of athletes, and whether Boston should get a National League team to join our American League Red Sox (I had to look it up).

My question stuck out like a sore thumb. I asked if there was ever any conflict between his writing/reporting about a game or event, and his obvious love for the game. He mentioned that it does happen, most notably in the ‘86 World Series. It appears that the Red Sox were close to winning, and had things well in hand. The fan in him was rooting for the Red Sox to finally break their curse, but the journalist in him loved the story of the innovative new way the Sox would have to find to lose the series.

As he put it:
On one hand, I wanted ths Sox to win. To finally break the curse, to finally get what they perhaps deserve and certainly earned. And if they won, the next day’s story would be so easy to write: Red Sox Won! But if they lost, I could point out all the ways they’ve come close and lost before, how they somehow managed to pull defeat out of victory. Which would you rather write?

Lots of folks nodded… Of course the more interesting story is if they lost. And they did. And it was a great story. And the folks who were nodding suddenly stopped; they were too deep into being fans to see the value of the story.

Afterwards, I talked to him for thirty seconds, but I got out of the way quickly. There were real fans who wanted to talk to him, and though I appreciated the fact that he could turn a non-fan into a fan for an hour or so, its even better to keep real fans glued for that hour when they could, well, be watching sports.

When I left, a line of 10 or so people were patiently waiting to buy a book, chat for a moment, get an autograph. Yes, like people would for an athlete.

Comments?

* * *

 

So others do use Judoscript also... · 04/09/2004 11:12 AM, Tech

David has been an active participant on the Judoscript mailing list, and some of his recent posts have been patches to the source to speed up some internal Judoscript operations. He has also posted them in his blog, and they’ve been accepted into the source by James, the leader of the Judoscript project.

But even better is this hint of things to come from the Judoscript mailing list…

Thanks, David! I’ll incorporate your code in the next revision.

Yes I am aware that constantly and repeatedly looking for methods is very expensive. When we compile JudoScript into JVM bytecode, this situation has to be addressed.

James

So, like Groovy, soon Judoscript will be compiling to bytecode, with all the speed and improvements that offers.

Comments?

* * *

 

CSS: 3 rights make a left · 04/08/2004 12:03 PM, Tech

As part of learning to work with Textpattern, I’ve had to learn to work with CSS. Lots of new concepts for folks used to blank gifs and tables for layout.

One thing that keeps confusing me (and perhaps others) is the difference between the DIV and the SPAN tag.

A bit of hunting reveals that, basically, DIVs are block-Level, SPANs are inline.

This means that if you put two DIVs next to each other, they come out as seperate blocks of content:
[DIV]
[DIV]

but two SPANs next to each other come out as
[SPAN][SPAN]

Usually, accd to my readings, you use DIV to provide some overall formatting, and use SPANs nested within to have additional styles on specific text.

As it turns out, TIMTOWTDI. You can set DIVs to display inline, and set SPANs to display as blocks… but that just confuses things.

Comments?

* * *

 

R: They just left the interface out... · 04/06/2004 11:06 AM, Analysis

A nice mention of the problems with R vs. well, everyone else by Nate.

I’ve tried R a million times, and every time, I have to drop it. I spend too much time trying to figure out how to get R to do simple things, things that are a single command in SPSS or SAS.

It not just a matter of a GUI (or the lack thereof). Its that no one has written docs from the 3 main user types:

It does have docs for the standard open-source groups:

It also has the requisite “Smart jerks who you admire for their smarts even as you wince at their lack of civility”. Spend 5 minutes on either Dev or User list and you quickly discover a user community which dislikes novices intently. (Yes, that’s a broad, sweeping generalization, but… There are many nice folks on the Dev and User lists, but it only takes a few jerks to make a newbie want to go elsewhere. The jerks also seem to jump on posters faster than the nice people. Funny how that seems to happen so often in online forums…)

Note that I have not one complaint about the capabilities! R can do pretty much anything. Analyses I’ve never heard of have multiple implementations, a great learning tool. They say the journey is the fun part, so I guess I should enjoy my attempts to climb Mount R and get to the good stuff… but I don’t.

We really need a SPSS user’s guide to R or something like that. This would address the 3 groups above, giving

In fact, for almost every programming language/tool popular today, we benefit from the cookbook, the design pattern book, and the “how to get things done” FAQ file. I find all of this missing for R. What’s there is a great start, but its just not enough for lazy ol’ me.

And before anyone complains,

One more point: If I hear one more person say “Its Open Source, get over it. Devs provide their own docs for Devs, so Users should provide their own docs as well. That’s what Open Source is all about”. Love the philosophy, but that what kills so many Open Source projects: They don’t invite users in, so no one can suggest those special ideas and enhancements that make a project grand, and then the Devs run out of interest, or code the project into a narrow niche.

Look at the great ideas on SF.net which are now abandoned. Why did Perl and Python grow? Because you don’t need to know how to code the C behind Perl to use it. And, lots of hand-holding books and docs were made to let even novices get started.

I want R to become the tool we look to and say “Oh, you could try SPSS, but why would you?” Right now, I find myself saying that about R.

Comments? [2]

* * *

 

ALL and ANY in SQL · 04/05/2004 02:26 PM, Database

There are some SQL commands which don’t get the daylight they deserve. Tom Kyte, of Ask Tom fame, always has good things to teach.

I enjoyed reading this entry which started off as a top-N query, and veered in all sorts of ntile and other directions.

Along the way, Tom pointed out that we often forget about the ANY/SOME and ALL options.

An example:


   select sal, deptno
   from emp
   where sal >= ANY ( select sal
                      from (select sal
                            from emp
                            order by sal desc)
                      where rownum <= 2 )
   order by deptno, sal

In this case, the query is looking for salaries which are greater than ANY of the items returned in the subquery. Why use the ANY? Because when you use >= with a subquery, it expects a SINGLE value, and a SINGLE row. If you don’t need a correlated subquery, this can be a clean way to code exactly the constraint you are looking for.

ALL works the same way, but means the condition must hold true for all of the returned values.

Yes, you could simulate this with a Max/Group by query in the sub, but this reads better in many ways.

Comments? [1]

* * *

 

SPSS allows DELETE VARIABLE now... · 04/02/2004 01:12 PM, Analysis

Another v12 feature: you can now DELETE VARIABLE. Another “why did it take 12 versions to offer this feature” feature. I know all those “add files file = * /drop = varlist.” folks will say its not needed, but which reads better to the novice?

Comments?

* * *

 

Jay Allen on the "value" of restricting RSS · 04/02/2004 11:54 AM, MetaBlog

Ok, not entirely the same thing as I refer to previously, but clearly on the same track.

Jay points out that if one chooses not to offer a feed, rather than encourage visitation, you encourage abandonment: If a site (or regularly updated part of a site) doesn’t haven’t an RSS feed, they essentially don’t exist to me.

Comments?

* * *

 

MIT OpenCourseware is amazing. · 04/02/2004 10:57 AM, Analysis

MIT has, like many times before, taken that next step. The MIT OpenCourseware project is basically a CMS system for professors to load up their course content… and its open to the world.

That’s right, the same courses on cutting edge physics, on the latest in numerical analysis, even the Sloan School of Management courses, all there.

Now, not every course is as complete as others: Some include no real content, just the “structure”. You would still need to buy your own textbook, for example, or find the articles referenced in the syllabus. But the infamous problem sets, exams, calendars, references, all the pieces you need to get “learned up” on a sophisticated subject are there.

Now, having met lots of MIT people, I can safely say that this in no way replaces an MIT experience. Its a unique place. But if you have no way to get anywhere near MIT, you can at least get a flavor of the knowledge they posess and build.

This is an amazing gift to the world.

And yes, its been around for over a year now… but with 200 additional courses added recently, and now “part I and part II” pieces of year long courses available, its worth a new look.

If nothing else, at least look at the Information and Entropy course. Unique in the world, this is an amazing way to view information. (Search on “information theory” if you dig this stuff for more info…)

Comments?

* * *

 

Tim is the Manns... · 04/01/2004 01:27 PM, Analysis

If you work with SPSS Clementine for long, you will run into Tim Manns. He has been with SPSS for over 5 years handling the toughest Clementine support issues. He now has his own site, http://www.timmanns.com/ and is now sharing some completely unsupported but very useful little apps, including some VB apps to work with Clementine-created PMML.

Another cool offering are XSL stylesheets to turn model-generated PMML into HTML! Very handy.

A nit: Tim, your javascript menus do not work properly on Firefox v0.8.

Comments?

* * *

 

Visual Basic Data Mining · 04/01/2004 01:20 PM, Analysis

Stumbled across Visual Basic Data Mining. Who would mine using VB, you ask? Lots of folks, and for good reason.

BTW, much of the same can be said for JudoScript since it leverages the JVM.

Some great stuff here, including how to work with PMML, product announcements, and good discussion forums.

They also have a blog with some very useful articles.

Comments?

* * *

 

On a previous episode...

Admin
powered by Textpattern 4.0.4 (r1956)