Individual Plus/Minus numbers

Solid effort by Morrissey in this game. He’s the one that got the team going in the second half with a couple of nice defensive plays (taking the charge under the basket and then diving on the floor to get a jump ball tie up) that were followed by a couple of quick baskets (a fast break off a steal by Battle and a spot up three ball on a semi break).

Pairs (plus minus numbers)

This bar chart shows the complete set of the plus/minus numbers.
Select any individual player on the left to see just his pairings (the bar where he is paired with himself represents his individual +/- number for the game.
Ctl-shft to select any combination of players to compare.

Pairs (minutes played)

Pairs (+/- per minute numbers)

The last game before conference play starts

• Penn State 74 Lehigh 55

Note: My prediction algorithm is based on Ken Pomeroy offensive and defensive efficiency ratings. I use my own methodology to tweak his ratings based on where the game is being played, then calculate a predicted score using the adjusted efficiency ratings and the expected pace of the game.

I present ULuv - because it’s all about u.

One of the things that I was looking to do with this site was to find ways to present the tons of data that tempo-free stats can generate in more interesting manners. Just looking at chart after chart of numbers can get quite boring.

The answer - welcome to the world of data visualization.

Visualization, as defined by Wikipedia, is any technique for creating images, diagrams, or animations to communicate a message. While computer imaging techniques may bring the tools to the masses, visualization has been around since the early civilization (heck, a hieroglyphic is nothing more than a visualization). To me, while KenPom.com and Basketball State websites provide invaluable information, they sometimes come up short on presenting them in attractive and meaningful forms (although the Basketball State does try hard).

The Big Ten Wonk took a step in the right direction last year when he used aerial charts (see below) to look at the efficiency ratings for the Big Ten Schools. The Wonk acknowledges being inspired by Iowa Hawkeye blogger Ryan Kobliska.

A step in the right direction but still with lots of room for improvement.

I tried to improve on that, at least from a presentation perspective not necessarily from a content perspective, by using the Google Chart API and Klowt to create graphics like this one.

At least to me, that chart is more aesthetically appealing.

Starting today, I’m going to take visualization to another level by creating interactive charts using tools provided by Many Eyes, an organization based in an IBM research lab. You, the reader, will be able to interactively control the content and layout of many charts that I use from now on. With “many eyes” looking at the data from different angles, there will undoubtedly be some new relationships uncovered.

Here’s a live example of what you’ll often see from here on out. I’ve taken some of the data I used in the “Who’s the best player in the B10?” series and uploaded it to the Many Eyes site. Here’s a “visualization” of that data in a bar chart form.

Once you click on the “Click to Interact” item, you see a bar chart of the points per game scoring average of the 30 players I looked at in my analysis. That’s interesting you say, but it’s just a bar chart and not a particularly interesting one at that. Ah but the fun has just begun.

Notice the legend along the y-axis of the chart. Select anyone of those items and you will immediately be presented with a bar chart of that criteria.

Don’t like the order of the items along the x-axis? There doesn’t seem to be any order to them (there actually is though - they happen to be ordered by minutes per game and I should have labeled the chart better - I’ll learn). Click on the Labels button on the left side of the x-axis and the bars will be ordered alphabetically by player (click once for descending order, twice for ascending).

Don’t like that order particularly either? Click on the values button and the bars will be arranged by their height (ascending or descending, your choice).

Want the original order back? Hit the data order button.

Do you find the data hard to read? Put your cursor on a bar and you’ll get a pop-up the details the underlying data.

Another feature. Remember the question that PSU76 raised about my methodology perhaps favoring guards? It got me thinking a little. While that specific question can’t be answered from this chart since I didn’t include any of the rankings in it. I did make a slight addition to my raw data by identifying the position that each player occupies. With that now part of the data set, we can do some additional analysis. There’s a drop down box for the x-axis at the bottom of the chart. Select Position from the drop down rather than Player and all of a sudden, all of the charts that were previously available to you for the players are now available as averages by position.

Another feature. Want to compare two categories. Simply hit ctrl-click as you select additional categories.

My one chart is now turned into dozens and dozens of charts that can easily selected you the reader of my blog. Play around with it and let me know what you think.

You can expect LOTS of similar charts in the future.

Hope you find ULuv as worthwhile as I think you will.