Color Theory Squeezed through the Google-ifier

The other day I was reading Miranda Mulligan's article on the subtle messages conveyed through map color choice, since we seem to associate certain attitudes with certain colors.
I figured that I could at least do a dry run of testing some of the associations I've heard before (orange is associated with appetite was something I heard years ago, and the more tenuous sounding association of purple with a sense of royalty -which just sounds like a regional cultural acquaintance).
So I fired up the Google and did an image search for lots of different terms to see how the interwebs have decided to associate images (which are just bunches of colored squares) with words.  Here's what I got, which is more color practice than color theory, but an interesting tool nonetheless...

Scooped!

Then, as I was typing this VERY POST, Edward Tufte named the snapshot results of Google image searches "quilting," which is a pretty clever name.  So you could say that these color tiles are a derivative of quilting.

Process
1) Google image search a term, screenshot the results (now called "quilting").  Each term generates quilt of roughly 1,000 images. 2) Blur and saturate for basic mottling of colors and indication of contrast, 3) then a full averaging of all those Google image thumbnails (around 2 million pixels) into a single, average, color (you'll have to push the saturation to pull out all the gray and isolate only hue -I went 80%).
Not sure how to average all the colors?  Neither was I, then I realized I could punt that to the image processor I was using (Fireworks) by scaling the quilt down to one pixel then scaling it back up.  I suppose you could get very slightly different results depending on the program (and therefore re-sampling method used). Now you know all my secrets.

Pro Tip: if you are going to try this out, or even if you just want to do some Tufte quilting and stop there, your life will be so much easier if you change your browser's zoom to the smallest possible percentage so you can cram in more tiny images in one go rather than stitch together a bunch of larger screenshots.

More, Cooler Stuff

I've broken out some of these into individual term/color images which are over at our Flickr photostream.  Also, because of the localization of Google search results, you can spoof other locations in the world to compare the cultural variability of color/term associations! The head spins.

Anyways, I have done enough painting in my life to know that the universe wants brown and the tendency of mixing a bunch of colors together will lead towards grayish brown.  That's why I was surprised to see so many deviations from that neutral.  An interesting follow up for someone with more chops than me would be to determine the extent and direction of each color tile's deviation from a web image "average" to score how how strongly the color/term association is -except when the association actually is brown.

Tweet Follow @JohnNelsonIDV Follow @IDVSolutions

Tourney Topo

Here is a really robust decision support system that will determine, for the geometric wonks among us, who to root for in the 2013 NCAA Men's Basketball Tournament, long after our actual favorites may have caught the bus.  Keep checking back for your updated cheering marching orders.

Click here to go full-screen.

If you are anything like me, proximity plays a significant role when determining who else to cheer on.  That, followed closely by logo coolness.  Anyways, here is a massively complex, super computing, hyper aware engine that considers proximity, and only proximity, in it's recommendation for your solely-spatially-fueled tournament support.

Thanks to Daniel Briggs, for his deft coding over a period of 2 hours.  And thanks to OpenStreetMap, D3, and Leaflet!


Tweet Follow @JohnNelsonIDV Follow @IDVSolutions

ISC West 2013

Session notes and resources for the ISC West presentation on April 10, 2013...

Title:
Social Media and Publicly Available Data: Turning Tweets and Other Data into Actionable Security Intelligence

Description:
How can publicly available data contribute to greater situational awareness? How can public- and private-sector organizations leverage information in social media to identify risks and mitigate threats from weather, natural disasters, terrorism or civil unrest?
An increasingly broad and deep array of publicly available information – active fires, vacant housing, hurricanes, earthquakes, census data and more – is available to risk and security operations teams. Learn to use this free data to improve planning, day-to-day operations and emergency response. Join data visualization expert John Nelson as he explores the new frontier of publicly available information – social media – and how to leverage information contained in tweets.

Helpful References:

• Click here for the "Anatomy of a Tweet" diagram.
• Signals of a movement via social data.
• More reading on the set of risk maps using public data.
• How-to make your own maps directly in Excel, with no add-ons.

Free Data Resources:

• data.gov a host of geographic and non-geographic data available from a variety of federal organizations in a variety of freely downloadable formats.
• Natural Earth A generous set of political and physical map files for generating your own maps.
• US Census Bureau Untold hoards of geographic and economic location data.
• Twitter API Get your people on this.
• Sleuthing Search Engine: “_____ data download”
• Diva-GIS A good source for international data.
• Try this with lots of spreadsheet data you’ve already got.

Tools:

• CloudMade A reliable and inexpensive means of creating your own basemaps.
• Tile Mill Another great tool for defining your own visual rules for basemaps and generating custom tile sets.
• OpenStreetMap A free worldwide map, with a global street network editable by you!
• USGS National Map A rich set of basemap imagery and services.
• UXBlog Check back here from time to time to see ginormous visualizations made with public data and other interesting projects we may be up to here at IDV Solutions.

A Globe Hack and Good Old Mercator

I just looked at the globe sitting on my desk and realized I still had pennies glued to it from a couple of years ago when I used it as an aid to understanding the nature of Web Mercator's scale distortion.

The globe was there, the pennies were there. The Tissot Indicatrix file was somewhere else. Where's the glue?!

These pennies are the analog version of Tissot circles, which are imaginary islands that are perfectly circular and equally sized -a super clever way of illustrating the various ways that map projections mess with reality. This globe and a pile of pennies were the easiest and most portable means I had at the time.
Compare the pennies on the globe with their approximations in Web Mercator. The penny island way up near the North Pole is the same penny you see down around Central America, but look at the resulting size warp! You might notice, too, that the pennies get increasingly bulgy at their top and start looking teardrop-shaped.  That is because over the north-south distance of the penny island the scale factor has increased that much. In fact, as you approach the North Pole the vertical stretching approaches infinity.

The imaginary islands that look a lot like pennies. Check out what Web Mercator has done to them!

Shoot, while I'm at it let me illustrate Great Circle lines real quick. Here's me snapping a straight line between two locations on the surface of the Earth. If I could fly and was in a hurry, this is the path I would take. But actual straight lines arc in the Mercator projection.  Check it out:

That photograph above shows me snapping a straight line on the Earth. When I warp that image into the Mercator projection you get a sense of how actual straight lines bend in Mercator (and vice-versa).

Web Mercator is becoming such a de-facto projection that I'm pretty sure that our zeitgeisty collective thinks of it as the actual shape of the earth. Yeah yeah, I know it's passé to pick on Mercator, and I suppose I am a bit, but it's important we keep this understanding of scale trade-off rattling around. So behold these globe-hack illustrations!
Shoot, Gerardus, forgive me; I am indebted to your mathematical and entrepreneurial skills. I can't stay mad at you.


Tweet Follow @JohnNelsonIDV Follow @IDVSolutions

P.S. I used to think that Mercator "maintained straight lines"; that straight lines in real life were straight in Mercator. But they are not -I misunderstood that so many years ago in my foolish misspent youth. What they really let navigators do, and why the projection was such a hit in the 1500's, was let sailors draw straight lines on the map which represented rhumb lines in real life. Rhumb lines aren't actually straight but as long as you sailed at the same angle on the ship's magnetic compass, you'd eventually get where you wanted to go (which was way better than dead reckoning). Here's more on rhumb lines, otherwise awesomely known to Mad Maxians as loxodromes.

Keep your bearing! Full sail! Source: Wikipedia.

Tribute to a Geographer

If you are reading this then it stands to reason that you are somehow interested in geography and/or design. As such, you should know that this person lived. She was my jr. high art and geography teacher. And she was my Mom.
If you find yourself near the Straights of Mackinac this June 22, and you are inclined, feel free to join us at 45.78386, -84.75724 to celebrate a life generously lived.

Chris Ann Nelson, 1949 - 2013

You taught me to look and keep looking for the loveliness in everything and every one. Through all of your peaks and troughs that seemed so scary at the time, I don't think there was ever a time when I wasn't fully aware of your love or confident of your support. You filled my head with the crazy idea that I could do just about anything and I lived to make you proud, right up until all I could do was drip water into your mouth and whisper that it was me who was proud of you and it was ok to let go. I'll miss you so.