Map of Where Tourists Take Pictures

Eric Fischer is doing work on comparing locals and tourists and where they photograph based on big Flickr data. It’s fascinating to try to identify cities from the thumbnails in his “Locals and Tourists” set. (I admit, I got very few right, either from “one at a time” or by looking for cities I know.)

Seattle Photographers

This reminds me a lot of Steve Coast’s work on Open Street Map, which I blogged about in “Map of London.” It’s fascinating to watch the implicit maps and the differences emerge from the location data in photos.

Via Data Mining blog and

The Future of Education is Chaotic, Fun and Unevenly Distributed

After I wrote “The future of education is chaotic and fun“, I came across “The Montessori Mafia” about the unusual levels of successfulness that Montessori produces.

In my post, I opened discussing how our current system of funding education in the US is to force everything through a government department. That department is constrained by a number of things, including “regulatory” capture by those parents with the time, inclination and skills to engage with the bureaucratic system. It’s also constrained by federal, state and probably municipal laws about what it can teach.

It’s also constrained by an instinctual conservativeness in parents, who think that they turned out ok, and so what worked for them will work for their kids.

But as the Montessori research demonstrates, all of those levels of conservativeness add up to a remarkable degree of sclerosis for the educational system. Now it may be that there’s good reasons to not adopt Montessori ideas for the general schools, just like there may be good reasons to not adopt Khan’s idea’s. But we’ve had a long time to study Montessori, and it seems to produce students who do well in life.

It’s not financial; the methods are old enough that anyone can use them, even if the early writings are probably still copyrighted.

So why aren’t its methods better distributed?

Elevation of Privilege news

I wanted to let people know that Microsoft is making the source files for the Elevation of Privilege game available. They are Adobe Illustrator and InDesign files, and are now on the EoP download site. They’re the 85mb of zipped goodness. They can be used under the same Creative Commons Attribution 3.0 US license under which we released the game.

If you’re not familiar with it, Elevation of Privilege is the easy way to get started threat modeling, and you can read about it here.

"Pirate my books, please"

Science fiction author Walter John Williams wants to get his out of print work online so you can read it:

To this end, I embarked upon a Cunning Plan. I discovered that my work had been pirated, and was available for free on BitTorrent sites located in the many outlaw server dens of former Marxist countries. So I downloaded my own work from thence with the intention of saving the work of scanning my books— I figured I’d let the pirates do the work, and steal from them. While this seemed karmically sound, there proved a couple problems.

Read more in “Crowdsource, Please.”

A Few Data Points

First, for those who might have missed it, Google has released Google Refine, a free tool for cleaning dirty data sets.  It allows you to pull in disparate data, then organize and clean it for consistency.

Next, some interesting thoughts on how “anonymized” data sets aren’t, and some thoughts on the implications of this from a risk perspective.  None of this is groundbreaking, but it’s nice to see some sane thinking about two facts that aren’t going away, no matter how much people might like them:  that data will continue to be accumulated and that it will be shared with varying levels of consideration for the risks of doing so.

Finally, yet another real-world example of risk homeostasis at work:  People who take vitamins make poorer health decisions in other areas.  Based on the number of times I’ve been asked questions along the lines of, “I don’t need to worry about x because I’ve {patched|installed anti-virus|switched to Apple|etc.}, right?” I’d say this still holds true for computing, too.

Now if you’ll excuse me, I have to go clean my anonymized data set so I can share it far and wide, which is OK since I’m going to encrypt it before I send it, right?


Photoblogging CHI2011

Last week, I had the pleasure of attending the ACM conference on Computer Human Interaction, CHI. As I mentioned in a work blog post, “Adding Usable Security to the SDL,” I’m now focused on usable security issues at work. I’m planning to say more about the conference in a little bit, but for right now, wanted to share my photographic notes. So here’s a Flickr set of pictures of some of the interesting talks I attended and posters I saw: “SIGCHI 2011 photos.”

Heaven Forbid the New York Times include Atheists

In “Is Your Religion Your Financial Destiny?,” the New York Times presents the following chart of income versus religion:
NYTimes Religion Income Graph

Note that it doesn’t include the non-religious, which one might think an interesting group as a control. Now, you might think that’s because the non-religious aren’t in the data set. But you’d be wrong. In the data set are atheists, agnostics and “nothing in particular.” That last includes 6.3% of the population as “secular unaffiliated” and another 5.8% as “religious unaffiliated.” Now, 6.3% is more than all non-Christian religions combined. Many of those non-Christian religions are shown in the graphic. Athiest, at 1.6%, is almost as large as Jewish, a major focus of the article, and 4 times larger than Hindus.

Now, you might also argue that athiests were left out because there were too few in the sample (as opposed to demographic data.) But there were 439 athiests, and 251 reform Jews.

Chris Wyspoal pointed out that atheists land after Hindus and Jews for 75k+ incomes.

All the news that’s fit to print, indeed.

Representative Bono-Mack on the Sony Hack

There’s a very interesting discussion on C-SPAN about the consumer’s right to know about breaches and how the individual is best positioned to decide how to react. “Representative Bono Mack Gives Details on Proposed Data Theft Bill.”

I’m glad to see how the debate is maturing, and how no one bothered with some of the silly arguments we’ve heard in the past.

The future of education is chaotic and fun

Lately, I’ve seen three interesting bits on the future of education, and I wanted to share some thoughts on what they mean. The first is a quickie by Don Boudreaux at Cafe Hayek, titled “Grocery School.” It starts “Suppose that we were supplied with groceries in same way that we are supplied with K-12 education.” It’s a fun thought experiment. The second is this video, which is worth watching through:

What Khan says about why his cousins prefer him on Youtube rings true. But “rings true” to Adam doesn’t matter here. By deploying the system online, it becomes a testable proposition. That will feed a level of experimentation and improvement that we haven’t seen in education. Before I get to what that all means, one more data point, which is this press release from the Gates Foundation, which covers how they’re spending 7.5 million on games for education.

To bring all of this together, what we’re seeing is recognition of the failure of the “Taylorist” system of education, where we think of education as a mechanistic transfer of information from teacher to student that happens at a constant rate, and the realization that we need systems that handle the amazing diversity of students. To do that, we’re going to need not just experiments followed by reform, but a constant process of experiment and adjustment. If you think the home-schooling movement and the early college movement have transformed school, just wait. Those are like Martin Luther’s theses, and they’re going to kick off layers of transformation. And with that transformation and chaos all sorts of things will emerge. A lot of them are good, but before we get there, there are several risks that we can foresee.

The first is that learning becomes but a set of discrete activities, rather than a lifelong process. That “hierarchy” that Khan discusses is not just a breakdown of learning, it’s inherently an exclusion of something that doesn’t make the chart. For example, research methods for young students will likely be a combination of search engines and critical thinking about web sites, with an ever-diminishing value placed on going to the library. You know, “reflecting the modern world.” Going to the library and really digging for an old book won’t help you stay at the top, and even a great teacher will only be allowed to award a few discretionary points. Another way to think about that is with the addition of game mechanics (the “leaderboard and badges” that Khan discusses) and games for learning that Gates is funding elsewhere, we’ll see a distinct quantization of learning. That the goal is to hit the badge, or top the board. When things become easy to measure, the hard to measure gets ignored.

Another risk is what will happen when teachers can see every moment of goofing off? Will that be used to drive diagnoses of ADHD even higher for normal boys? Will “not living up to potential” be a new and ‘data-driven’ part of the report card?

Yet another risky area is privacy and commercialization. Will digitization lead to marketing of new tutoring systems? Will report cards and transcripts convert from an A-F summary to a computer-readable XML-encapsulated explanation of every problem set that little Robert and Jane have been through? (And note that that the process of making the transcript computer-readable will further drive out non-standardized activities. If there’s no code for “went to the rare book room,” there’s no credit for it, and if there’s no credit for it, why bother in today’s hyper-competitive world of college admissions?)

Interestingly, one of the few broad privacy laws in the US, the COPPA protects those under 13 to some extent, but I don’t think it considers educators as much as commercial web sites, and when Harvard demands the XML file, well, you can opt-in to their admissions process or not.

So even recognizing those risks, we’ll likely get an educational system that will stay with students who are having trouble without holding back those who want to move faster. We’ll learn to give students skills and approaches faster and faster. We’ll have to figure out now to ensure students learn cooperation, project management and other harder to quantify sorts of things. But I do think that we can give kids more skills and knowledge faster and better than we do today. If we do that, the world will have more smarter people than ever before, and even more interesting chaos will emerge.

I suspect there’s other things that will predictably go wrong, and other outcomes that I’m not seeing. What do you think?