ESRI Conference at PSU

So, last week, just before I caught the plague, I was lucky enough to attend the ArcGIS 10 user conference held on the Penn State main campus.  We arrived fashionably late, sat in the back of the ballroom, and were therefore first in line for the buffet.  I became violently ill that very evening, so I would like to take this opportunity to publicly apologize to the 150 people who touched the salad tongs after me.  Also, so sorry for anyone who voted at the far left voting booth at the Penn Township Municipal Building after 4pm.  Ahem.

There was no name tag printed for me, and I have to say, my one regret about the day is that I lacked the courage to steal another person's name tag.  Said person registered for the conference and did not show up.  Said person was named "JingJing Wang".  Holy God.  I think JingJing Wang was too busy being awesome to attend.  And so nameless, I attended the conference.  Damn.  I can't believe I was too chicken to become JingJing Wang.  I guess I'm not ready to be that awesome.

Really, the main thing that I got from the conference, to distill down to a nice blog-bite, is that ArcGIS Online is pretty freaking rad.  Sure, it's no ArcGIS 10, or anything, but then again it's free.  Just make an account (I am ArnoldJudasRimmer*, because I am a colossal geek) and you can make, post, share, and edit maps online.  For FREE.


View Larger Map

Above is a super simple map of my old neighborhood with a subway overlay that took me about 2 minutes to make.  Just zoom to location, add basemaps (same ones we use with ArcGIS 10) and any layers you would like to add, from ArcGIS Online or elsewhere.  Simple and easy to share.  I'd really recommend starting an account and playing around with the website.  It's pretty nice.

And you can also make presentations.  Here is a particularly nice one by "bszukalski" showing more of what can be done using ArcGIS Online.

View Larger Map, click "Start Presentation" to view






*

Lister and Rimmer
"It's cold outside, there's no kind of atmosphere..."

If I've learned nothing else from this course, it's that Mercator sucks out loud.

Stop.  Using.  Mercator projections.  Just stop.

Because this is not getting old anytime soon, here is a great illustration that was linked on one of my very favorite sites, FlowingData.com.  Kai Krause has taken a goodly part of the world and superimposed it onto Africa, with very little overhang, just to make the point that Africa is freaking huge.  Observe.

yes, it is that large.  Click to make even larger.

By country, if you prefer.  Click to enlarge

The linked blog above, which I cropped these images out of, has a nice sidebar with actual country sizes, just in case you want to look them up.

FlowingData makes a point of showing the GoogleMaps Mercator projection to show how ridiculously inaccurate it is. I found a really lovely satellite image of Earth in standard Mercator projection (really lovely.  just like all those cheesy PSAs around Earth Day say- it is an amazingly beautiful planet.  Jersey notwithstanding.) and cut-and-pasted roughly similar-sized rectangles side-by-side of Greenland, Africa, and much of North America.  Behold.

Greenland:  "Where is your God now, mortal?"

I, for one, would like to welcome our new overlords from the far North.  Herring for everyone!

Looking at the true size of Africa, can one really wonder why all those pasty-skinned Europeans ever looked south and thought, surely they can't be using all that land, can they? * 

*Kidding, kidding, colonialism was bad.  Especially for Congo.  Damn.

Olympic Medals Assignment

Here is my map for the Olympic assignment, click to enlarge

Well.  We were given the assignment to make a map using a proportional symbol scheme to show which countries won the most Olympic medals at the Beijing 2008 games.  Simple enough.  Almost too simple.  I have to say, I am not crazy about my map.  I kind of phoned this one in.  And don't worry, I would also tell that to my professor (who, BTW, reads our blogs) so this is not something I would not say in class.  My map is okay.  But merely okay.  It is not superlative.  However, I have been sick as a dog this week (I even missed class, which by and large Does Not Happen) and this was the best effort I could muster.  I then went home and slept for ten hours.  

So.  Why all the bitching?  Well, I've decided to present this as a post-mortem, and to detail what went wrong and what could have been better.  Including one detail that I will attempt to fix really fast before class.  Really fast.

What I don't like:

• My medal symbol.  I attempted to import a really nifty image:

Click to enlarge, here is the original

But I had to take it into GIMP and remove the background (which took goddamn forever), and then when I imported into ArcGIS as a PNG for some damn reason, I wound up with a pixelated, yellow-toned, piece of crap.  Eventually I said F--- This and decided to make my own damn symbol, even if it looked less awesome.  So I made a hollow circle to have a sense of openness in the event of overlap but with a center dot to give a notion of placement, so that you wouldn't think that say, Austria's symbol was over Hungary.  I chose orange because... it showed up nicely.

Wow.  There is a lot of damn white on that map.  I would have liked to tone it down by giving it a slightly toned background, like a beigey shade... But my Olympic ring symbol, which I de-backgrounded in PP, decided to import in with a damn white background.  I decided that I couldn't deal with the sinus headache anymore, and just kept a white background.

The text in the bottom right should be smaller.  I honestly had trouble keeping track of the actual size of my text, as this map is drafted to be 22x34.  I may touch that up before class.  Maybe.

The north arrow does not curve into the side of the earth.  I just stuck it there.  Hrm.

The projection (Winkel II:  The Quickening) that I chose I liked initially for the pleasing shape that it gave the Earth.  However, looking at my map, I feel like my data is sort of not front and center.  Does my viewer have to squint and sort out which are the top medal winners?  I almost feel I should have chosen a scheme that unfairly emphasized the Northern Hemisphere to hype up my data.  Sorry, Southern Hemisphere.  First colonialism and now this.

Hey!  Let's all look at Antarctica!  There sure is a lot of it, doing f---ing nothing and taking up valuable real estate.

So I inserted a little box to show the top 10 medal winners.  Pretty much because my map is really bare and dull.  And later, when I look at my map, I think, where the hell is Germany?  Oh, GD ArcGIS ate my label with an overlap.  While I'm sure the French are glad to have some payback for WWII, I'm not too happy about my label.  I'm going to try and fix that Monday.

I made the dick move of deciding to only label countries that scored in excess of 20 medals, in the interest of not cluttering my map.  My apologies to Africa, South America, and Oceania.  Oh, and to Palau, partly because of my map but mostly because of global warming.

The projection is a little silly, in the way that all projections are silly, because it is inherently false while attempting to be realistic.  Sure, let's be all forward thinking and project the world in a round-ish shape... and then be all fake and cram all the continents to the front in a way that, God willing, they will never be.  Looking at the earth that way reminds me of the way actors sit at a table in a television show or staged play, all crammed together but avoiding one side, in order to all be visible to the audience at once.  Real people don't eat dinner that way.  Real continents don't look that way.  But, well, artifice it is.

***************UPDATE************************
It's twenty minutes to class.  I came in early and fixed/triaged my map.  Here it is.  Of course, I sit next to Foram and am in despair at seeing how f---ing awesome her map is.  Well.  My map is below.

well, it's done, anyway.

Map of Berkeley Pit in Butte, MT

We were recently given the class assignment to create an 11x17 map of the Berkeley Pit in Butte, MT.  Butte is a city that was essentially built up by, and eventually brought down by, the copper industry, and the Berkeley Pit is the legacy of an abandoned strip mine that has since filled with toxic groundwater.  It is now America's largest superfund site.  A flock of migrating birds died rather horribly following a stay at the Pit.

Our maps are required to show 1. a hillshade map, 2. an image of the Pit, 3.  a contour elevation map, 4.  some basic info on the city, the pit, and anything else we feel is necessary.  I was pretty happy with the way mine came out, so I am attaching an image of my map with a few notes on how/why/what I did.  Any critiques are welcome.

My map, woot woot, click to enlarge.

Okay.  So, first off, you can see my two created maps, hillshade and contour, as well as a 3D image that was captured in ArcScene (bottom right corner).  They are okay- I flat out do not like the hillshade map, as the resolution is not as fine as I am used to seeing from PAMAP DEMs.  The contour map is merely okay as well.  The design choice that I like about them, however, is just aesthetic-- I like the placement overall and the unifying orange shade that ties the whole piece together.  I used the eyedropper tool to select that shade from my retro top-left corner item and dropped it into pretty much every element of the map.  It was only later that I realized I had a very nice Halloween theme color map.  

My very favorite element in the map is the retro-dystopian image at the top-left corner.  I initially wanted a nice depressing touch for my nice depressing map, and hoped to get an image of a sign reading "Welcome to Butte" that was imbued with cheer while simultaneously radiating decrepitude.  If that is a word.  Is now.  Anyway... I found the great image of the happy miners, celebrating Butte's status as "The Richest Hill in the World!" and removed all the text in GIMP and put in my own depressing and sarcastic factoids using PowerPoint 2010.  Two birds, one stone-- a nice image that shows the decline of the city from a historic high, which also functions as my textual information.  Below is the original "happy miners" image, so that you can see the text as it was.  

"Land of the Shining Mountains", click to enlarge

Voronoi II-- Electric Boogaloo

So, as promised, more on Voronoi and GIS.  I found a great video on youtube, of a scholarly presentation on spatial query processing and Voronoi diagramming.  A young man, Mehdi Sharifzadeh, is presenting his research:



Don't worry, you don't have to watch the whole hour.  I have to say, I've seen _most of it (sorry, Mehdi, I'm a busy girl) and can assure you that the mathematics presented are not at all challenging, if you are math-phobic.  If you want to skip around a bit, I would recommend beginning to watch at about the 8 minute mark (nice discussion of spatial data querying), or beginning at the 11 minute mark, for a very good discussion of what is Voronoi (and ignore the comments from the audience by the guy who clearly does not understand basic geometry, and the lady who likes to hear herself talk).    MS does a lovely job of describing the Voronoi cell as "capturing the concept of closeness" and the points p as "a force forcing something outward".  That point of equal force could then come to define a boundary of each irregular polygon.  I found the video very informative.  Starting at about the 17 minute mark he gives a good example of Voronoi diagramming applied to GIS (with some variation).

And just in case you just haven't gotten enough Voronoi awesomeness, here is a totally trippy video clip:



The music is either from an old video game, or a really terrible low-budget corporate training video.

Damn.  Now I really want to watch Office Space again.

And, finally, I found a super-cool image of the United States rendered with each state as a Voronoi-generated polygon with each state capitol serving as point p.  My thanks to Scholars' Lab for the post, which I would recommend for you to read.  Click each image to imbiggen.  

Ordinary USA

Voronoi States of America!

And now in Stereo!

Now here's what I want to know-- how do you adapt a Voronoi GIS query to incorporate topography?  If anyone has an answer, or knows of some good places to look, please let me know.

Voronoi Diagramming and GIS

So in my entry on John Snow's cholera map, I accidentally stumbled upon the concept of voronoi diagramming.  I wanted to know more about voronoi, to understand it better, and have found some very interesting mathematical/geospatial websites and videos.  No, really- come back- it's really cool!

Okay.  So what is voronoi?  Here is more or less what I have absorbed, and my sincere apologies to any actual mathematicians in the crowd-- I am a moron.  Essentially picture a series of point data scattered on a page-- say, cities in the state of PA, or Duane Reade locations in NYC, or marbles scattered on a table top.  The voronoi diagram is generated by defining an irregular polygon around each point of data p whose border is defined by an equidistance between each other p, and therefore any point within the polygon will be closer to its own parent p point than any other point p.*  You will wind up with something fairly ugly, really, like this:

Voronoi Diagram

of course, in nature it's not so ugly.

Voronoi in dragonfly wing

Voronoi in turtle shell

To get the concept, just picture a dot in the center of each irregular polygon.  Voila!  Voronoi.

So, why am I posting this on a cartography blog?  Well, there are some pretty clear applications of Voronoi to GIS.  To go back to an earlier example I gave, Duane Reade locations in NYC (and there are ... about a million)...

You can't swing a dead cat in Midtown without hitting a Duane Reade, believe me.
View Larger Map

...picture a mapping query that allows you to block up NYC into polygons with data point centers p  that each represent a Duane Reade.  You can then take any other point on the island and immediately know its closest location to a Duane Reade, and route accordingly.

I found so much interesting material on Voronoi that I believe I will have to split this blog into two posts, and will have more to share tomorrow on Voronoi as it applies to GIS.  Until then....




*Another nice definition, if you prefer something a little cleaner than my idiot-savant distillation (because, really, you, like me, may need to read several different definitions before it really clicks), is: A Voronoi diagram divides the drawing into regions around each point that are shaped so that the borders of the regions are equidistant from the two nearest points. It's kind of the vector equivalent of grid mapping where you're trying to show boundaries of influence from point data.

John Snow, Cartographic Breakthroughs, and the Cholera Epidemic of 1854.

Recently I began to reread one of my favorite historic science books, Stephen Johnson's The Ghost Map, which deals with the great cholera epidemic of 1854 in London.  At the time of the contagion, cholera was believed to be caused by miasma*.  Fieldwork by a pioneering doctor named John Snow established a crucial connection of the cholera victims-- they had all frequented the Broad Street pump for their daily drinking water.  The well of the pump itself was infected with cholera bacteria.  Snow's door-to-door surveying work established the cause of the contagion, as well as the notable fact that cholera is water- rather than air-borne.

No, no, cartoons are supposed to be funny.

Snow's work is to great interest to cartographers as his resulting map provided particularly damning evidence that the pump was the source of the disease.  He conducted street level research and found that although certain other pumps were closer to victims as the crow flew, the Broad Street pump was usually the closest to the victims based on street traffic paths.  His published map is below:

Snow map 1854, click to enlarge

Snow also stumbled upon voronoi style mapping, and below is his famous map with the voronoi region overlay:

Voronoi map, click to enlarge

*definition of miasma (noun):  a mysterious mist that killed thousands before it vanished entirely in the face of modern science, never to be seen again.