Spatial Analysis - starting to work with some new tools

Once again, I feel like I am starting over with a whole new set of learning in this class, which is a great thing (however daunting it may be!).  In returning to the hurricane and tornado data, I realized, first and foremost, that my naming of my various map layers left a lot to be desired, and I had to spend more time that one would like with various layers' attribute tables to figure out what various layers were in fact representing.  Despite that annoyance (a lesson learned, really), I decided to try to figure out if the "mean center" of the tornado lines that we had created a few weeks ago changed over the decades.  The idea for this approach came from seeing some of my classmates' blogs that suggested both a westward shift in tornadoes over the years as well as a counter-argument that tornadoes' locations were not shifting significantly. To create such maps, I undertook a few steps:

1) I used definition queries to create lines that represented the tornado lines for each of three decades (the 1970s, the 1980s, and the 1990s) and changed the symbology to make each of these stand out.

2) I then ran the Mean Center analysis on each data set and matched the symbology of the mean center to the symbology for the decades analyzed.

Two thoughts: first, the mean center of the tornadoes does not seem to have shifted significantly across those decades, hovering throughout in southeastern Missouri.  Second, and far more importantly, I kept getting a warning that "The input feature class does not seem to contain projected data," which I learned through ArcGIS Pro means that I need to set a projection for the feature class.  The same issue I faced with that blasted historical map during the check-in!  Luckily, ArcGIS Pro also provided a solution for this issue, which I tried out.  After re-running the mean centers against these newly "projected" tornado lines, the resulting map is below.

We actually do see that correcting to include projections did change the mean center, so I am glad I took this extra step and learned how to create projections for feature layers in a relatively straightforward way.  My biggest takeaway from doing this little blog post, however, is to create a schedule to do significantly more practice with spatial analysis in the labs this week.  I'll keep updating the blog as I work through kernel density and other tools--feel free to revisit for riveting updates! (I know that, in reality, these are updates that not even a mother could love.) Onward.

Comments

  1. sam41803November 6, 2023 at 11:45 AM

    yay it worked on unmodified chrome!

    Interesting approach, using the Tornado lines rather than start/end points as I bet was common. The trickiness here is there's just a lot of tornadoes through the period, so the lines can look rather noisy... Not sure what would work formally under stats. Maybe you can do something on the length (like the length of the different vectors over time/area)?

    ReplyDelete
  2. Sara R.November 6, 2023 at 12:04 PM

    Woohoo I can comment! This looks really nice Mark! When I tried to do this assignment i had something like this but I had thought that I did it wrong so I deleted it. Now I see I was on the correct track.

    ReplyDelete
  3. Caroline HallDecember 12, 2023 at 12:06 PM

    This looks great! I tried to recreate the same type of analysis but somehow tornados were also showing up at the N0E0 location in the ocean next to Africa, so needless to say my mean points were quite off.

    ReplyDelete

Post a Comment

Popular posts from this blog

A "time series" of maps showing rail expansion across the northern Midwest--and the Swedes and Norwegians who followed the rails?

Image
The series of maps I would like to create still lies well beyond my skillset (a reality that is likely to be true for some time), but it was fun to use the "time" function on ArcGIS Pro to get a sense of how railroads were expanding across the country in the mid-19th century.  Given that my mother's paternal side came to the US in the late 19th century from Sweden, I decided to look at the Swedish and Norwegian populations in the 1880 census data, and it was no great surprise to see that Minnesota and the Dakotas had relatively large concentrations of these populations at the time.  This map brings together the Swedish and Norwegian population as normalized across the total population, railroads in 1880, and major rivers. A couple of things stood out to me looking at this map: first, the Swedes and Norwegians in this northern Midwest area were, indeed, concentrated along both rail lines and rivers--notably the Red River that runs south from Winnipeg (and eventually create...
Read more

Digitizing 1923 buildings in my hometown, Summerville, SC

Image
 I suspect that our blog posts this week are going to be a bit less analytical than those in previous weeks, given that we were largely learning how to georeference historical maps and digitize features.  Click, click, click, click.  Below, please find a georeferenced map of my hometown, Summerville, SC, in June 1923, with a few blocks of buildings digitized. There were several steps in the process that did surprise me as I georeferenced and digitized features.  First, I was not totally surprised that the blocks in downtown in 1923 more or less matched the blocks today, given that I chose the downtown area to georeference.  (There was even a little bend in the road on the bottom right corner of the map that is still there.)  I was surprised, however, to learn that the railroad tracks were in exactly the same place since I know that there have been many changes to the rail lines themselves across the SC Lowcountry.  There were some fun buildings to find...
Read more