I have been using some sort of spatial analysis tool since August 1992. It is easy to remember the precise time, as there was a certain ill-mannered gentleman that came to visit on the 24th of that month – Andrew. Hurricane Andrew, that is.
Just a few months earlier, our new Vice-president of Operations posed a valid question to the Research & Development department: “How many high-value homes do we have insured, and where are they?” Selecting the properties and counting them was just an SQL statement away, and a report by county was fairly easy to create, but there was no optimal way to sort 67 rows of data, and county-level resolution was not fine enough.
As we examined our options, we decided to try mapping some of the risks by Public Land Survey System (PLSS) coordinates. Being the Farm Bureau, we insure a decent amount of rural business. For many years we had been collecting section / township / range coordinates for our property business as an aid for driving to appointments with insureds. So now we have a way to aggregate data down to a specific grid cell (typically one square mile), but the problem of how to present the data at this resolution was quite impractical without a map.
Fortunately PLSS data lends itself to graphing by its nature, and Florida especially so as it does not have multiple meridians (dividing lines between N/S townships and E/W ranges). You certainly would not have attempted this in Ohio or Mississippi with all their meridians. With the understanding of how the corner monuments were marked off, the next step was to write a function to convert PLSS coordinates into standard x/y pairs with the hopes of being able to graph them. (For more information about PLSS, check out the Geography Department PLSS Lecture Notes from the University of Texas, or the PLSS Fact Sheet at WhiteStar's site.)
At that time, Lotus 1-2-3 v1.0 for Windows 3.11 was my main application. It had all the math functions necessary to convert the coordinates, AND the ability to graph the points. We made the decision that the upper-left-hand corner of each section was the control point. That meant that PLSS 006 01S 01E was at coordinates 0,0 on the x/y graph. Every Township North/South was a six mile positive/negative offset on the y-axis, and similarly each Range East/West was a six-mile positive/negative offset along the x-axis. Since the section numbers snake back and forth across a township/range block, we made a lookup table on section to make the remaining adjustments.
Soon we were able query exposure data off our mainframe, aggregate the data by PLSS coordinates, and display points on our map/graph within Lotus 1-2-3. The obvious problem was that points were all we had, making it quite difficult to interpret where the graphed points were located in relation to important landmarks (like the shoreline or county boundaries).
Okay, so now things really get crazy. Since we are able to graph PLSS coordinate points onto an X/Y graph, what is to keep us from drawing a connected series of points and drawing features?! All we had to do was collect the STR coordinates of the significant vertices of Florida counties. After testing the notion, we determined there needed to be gaps between some rows of data to simulate a pen lift from the last point of one county before starting to draw the next county. The results can be seen below.
Click to enlarge
If you really want to feel my pain, just think of the thematic maps I created with a screen capture of the above map and filling in the counties and legend by hand in Windows Paintbrush. Hours spent on something that I can do now in moments.
So, was all that effort worth the time? It sure was. When Hurricane Andrew struck, we obtained all the information we could about wind speeds superimposed that over our exposure, and estimated our damages by construction type. We were able to come up with an estimate that was only off by 2% from the ultimate payout.
Soon afterwards we received a brochure in the mail promoting a GIS platform. We had already seen the value of such a tool, and the amount of time it would save over what we were doing. Now there is hardly anything we do that does not include some degree of spatial analysis or presentation.
Unfortunately, I have not been reformed after all these years. Now that we have all this wonderful technology, I still find ways to abuse it. My passion for die cast cars has forced me to map every department store, toy store, drug store and mall across Florida. Whenever I have to leave town, my route is altered to pass through as many of those places as possible.
My custom crafted display case / coffee table
in my office full of die cast cars.
Yes, GIS does indeed enrich our lives.