Terrific starting-point for an investigation — brave, actually. I wouldn’t have expected that anyone would actually respond to your survey. Wow, collecting 71 pin numbers. Hard to believe, and very interesting. LOVE the HEAT MAP. Dope study.Please LABEL your diagram, since each of the rows means different things, which was only explained by your talk. You might want to talk to a stats expert to see if you can come up with some verifiably significant differences between men & women etc. 
This project suggests lots of other visualizations, eg. a 100×100 grid showing first 2 numbers against second pair. 

Really cool!  I love the heat map idea, too!  It’s looks very cool.  I kind of wish that the squares had labels (either below or mouseover) to make it easier to see the trends you talk about in your paragraph. Demographic filters would be cool too. Nice work! -Amanda

If you’re ok working for The Man, I feel like there are some people over at CERT/SEI who’d be more than happy to give work like this exposure (or augment your data) in some way…Very nice work. Blurs the line between research and art nicely. -SB

Do have to wonder what would happen if this study ‘got out’ and if hackers could use it to figure out some of the most likely pin numbers.

I’m interested in seeing almost a hybrid between this and the colors across cultures project. Does culture have an impact on the numbers you choose? I know some cultures place a lot of value on numerology eg. Chinese

I wonder how many people would answer this if they got paid for it on mechanical turk. Would they tell the truth or try to game the system? Even if they do lie it could be interesting. See Benford’s law and http://www.wnyc.org/shows/radiolab/episodes/2009/09/11. -MH
– but pin numbers might not apply to Benford’s law – after all, the only rule is that people have to make them up (both when choosing their PIN and when faking it).-Jon
– Listen to the radiolab episode anyway. Radiolab is great. 🙂

The number pad visualizations are a little hard to read.
The number pad visualizations are very difficult to read. Without any indicating text, you have a jumble of information. You can’t expect people to even know what it’s about if you don’t even put the title in the visual. You need to think more about how to show as much information as possible with the minimal amount of visual.
Some labelling would be helpful. Also a lighter background- the black background is washing out the subtle variations in the heat map colors.

Did you take account for banks that do not allow you to choose your PIN number.  How much do you think this would affect your stats if not?

Are there any number pads the reverse direction?