Lab #7: Cartography

This first map was the least surprising of the three: it shows the percent change in county population of "some other race"--in other words, people who didn't identify as Black, American Indian, Asian, or Pacific Islander--over the decade from 1990 to 2000. The states that have the highest percent changes are, the border states: California, New Mexico, Arizona, and Texas. The map shows -- unsurprisingly -- that most of our immigration is happening through the US-Mexico border. Maybe if the census had a table for people who identified as Latino/a for this data set, the map would not look the same. I do wonder about that little splotch of color in Washington: do Canadians just really love that county in Washington?

This second map I still have questions about. It shows the Asian Population Density per county in the year 2000. I expected there to be an obvious concentration in California, but what we instead see is areas of high density in California, but also several areas of high density across the South, Mid Atlantic, and North East. Maybe this is the consequence of spending all my time in the Bay Area and Los Angeles county, but I thought there were far fewer Asians in the South than California overall. If you look at the map, though, you'll see that the counties in the eastern US are much smaller than those of California. In order to calculate density, I divided Asian Population of each county by the area of each county, so perhaps county size is what's making the difference in this map.

Finally, this map shows Black Population per county in 2000. The counties with the highest populations seem to be Southern California and the North East, as well as the South and Mid Atlantic. Just like we saw in the map for Asian population density, very few Blacks live in Middle America. I'm not very familiar with the demographics of Southern California, so the high population of San Bernadino County, the large county east of Los Angeles was surprising. We have to remember  though that is a map of straight population numbers, and not population density, which could explain San Bernadino--one of the largest counties in the US.

It's amazing that the government provides so much raw data that anyone with mapping software can just play with. Data collection seems like the tricky part; once you have a nicely formatted table, you can do a lot to analyze it relatively easily. As far as the census goes, I was a little disappointed in the some the data they didn't have. "Black, White, Asian, American Indian, and Pacific Islander" is a pretty pathetic set of groupings for the United States population. What about Middle Eastern? What about Latino? Were these not included because they're classifying people by "race" instead of "ethnicity"? These are ambiguous theoretical questions, but they clearly make a difference when you're showing them on a map. I also wonder about break values: having a legend makes the maps above transparent, but the maps could be manipulated to look a totally different way if I changed the break values. The census tutorial hinted that there are standards for US Population values so I used those, but I wonder if many people ignore them to achieve their desired effect.

As always, I'm impressed with my overall GIS experience, but wish the focus of our tutorials was less on fine details ("this is how you change the specific colors) and more on concepts and broad skills ("this is how to calculate values in a table. try doing several of them.") It seems like the exercizes emphasize both about the same, and then I lose the important pieces amongst the fluffy fun stuff. I'd rather my conceptual understanding be solid.

Lab #6: DEMS in ArcGIS

The maps below visualize the Bay Area Peninsula, where I grew up. San Francisco is the sparkling, urban 7x7 mile area at the tip of peninsula. When people from the bay area say they live on "the peninsula," they typically mean one of the cities south of San Francisco. So as you will see in the maps below, the top of the peninsula is not included. Instead, the main thing you see is the hills of Golden Gate National Recreation Center (blue, in the first map below). The fog rolls over these hills from the west into the reservoir (brown), which is just east, nearly on the San Andreas Fault Line. East of the reservoir is Highway 280 (not shown), which runs north and south along the peninsula separating the beautiful landscape from the residential areas. Many of these areas are on much flatter ground, which is why the rightmost portion of the map directly below looks like there might be ocean where there is really flat land. Around all of this land, of course, is the San Francisco Bay (east) and the Pacific Ocean (west).

With the exception of the 3D model, which is tilted slightly to the right to give you a better view of the interesting landscape features, each map is oriented with north at the top. For reference, the geographic coordinate system for these maps is GCS North American 1983, and the extent information is as follows: Top: 37.71 degrees;  Bottom: 37.34 degrees; Left: -122.62 degrees; Right: -122.62 degrees.

Lab #5: Map Projections

Six map projections and their planar distances between Washington D.C and Kabul, Afghanistan.

My favorite example so far of the huge variation in map types and functions is the one Professor Sheng gave in lecture about BART. What are the important features on a BART map? Is it the distance between stations? The coordinates of each station? The shape of each station? It isn't any of those really. As a Bay Area native, I can say for certain that nobody on the Peninsula or the East Bay cares minutes and degrees when they're going into the city (save the degree of fog)--not even distance.

What is important is the relative location--the order. Which stop do I get off at? When the lovely Bay Area Transit designers were making the BART map, they probably didn't worry too much about whether the scale was spot on. There are plenty of instances, though, in which precision is much more important. Say you want to know the distance between New York and LA--or hey, Kabul and Washington D.C? How far apart are they? An equidistant map projection, like the two shown above, might give you a good number. Some other map projections distort the distance between locations in favor of preserving other qualities, like area (See Bonne and Eckhert IV) or shape and angles (See Hotine Oblique Mercator and Stereographic).

This has positives and negatives. On the one hand, we have a lot of map projections at our fingertips, increasing the likelihood that there is one out there that measures what we want it to measure. If you look at the Equidistant maps above, you'll see that they have two very different functions.The Azimuthal Equidistant map preserves distances from the poles outwards, while the Equidistant Cyllindrical projection preserves distances across meridians and equators. That is why the Azimuthal produces a distance of over 8,000 miles, while the Cyllindrical produces one closer to 5,000. It turns out that in the case of Kabul and Washington D.C., neither of these is perfect, although the Cyllindrical map comes closer, since Kabul and D.C are at similar latitudes.

The downside is that map projections can be misleading and manipulative. Distortion isn't always apolitical. During the Cold War, for example, the United States used Mercator projections because they distorted the size of the Soviet Union, making it monstrously larger than it really was. What the Mercator does is distort areas based on their distance from the equator. So Greenland, for example, is larger than Africa, even though Africa is at least 10 times as huge. Maps shape the way we think about the world. They can shape the way we feel. Africa is a good example. A lot of people don't realize that the United States can essentially fit inside the Sahara Desert alone. Maybe if we had fewer cheeseburgers and better map projections growing up, we wouldn't be such big-headed Americans.

Lab #4: Introducing ArcMap

My GIS Experience:

My overall experience working through the GIS tutorial was enjoyable. I like maps. I like data. I like I'm learning a skill that allows me to make data digestible and beautiful. Most of my frustrations with the tutorial came from the length and tedium.Instead of working through one long, complex mapping excersise, that covered everything we were meant to learn, it would have been nice to work with a smaller data set--maybe with a more familiar subject (what is a parcel? I still don't even know)--and learn how to use the tools one by one.

I also found myself wanting to understand the system deeper. Where does this data come from? What does it look like in its raw form? How do the different layers interact with each other? I can tell that if I understood some of the foundational technicalities better, I would have a better grasp of what I'm doing when I manipulate the visual representation of the data. But there's clearly a lot to learn. Maybe this is both a pitfall and a potential. GIS is tricky and hard to democratize, but once you learn the skill, you have a lot of power at your fingertips.

The other thing I wondered about was how users collaborate on maps. It seems like it would be so easy to mix files up or save a pathway wrong. If I were working on a serious project with a partner I would want some way to document our progress, as well as our sources. Where did the data set come from? Has information become so accessible that it's normal for GIS programmers to regularly give and receive data sets from other entities? That could be another issue in terms of trust and credibility.

I'm interested in what kind of software is available to do GIS on macs. Frankly, I'm not interested in honing a skill unless I can do it on my own laptop, and I'll probably take what I learn from this class to explore what the other options and technology there is out there.