Corridor & Cost Path Analysis

The cost of movement is usually something thought of in terms of money, in the context of shipment or transport.  Cost can take the form of a whole host of other factors, though, when it refers to some expenditure to move across a landscape.  We can express cost in terms of time, energy required, or even impact on the environment.  Anything that effects movement can be modeled as a cost, which can also be translated into a predictive measure, as in the case of corridor modeling and analysis.

  
The above is a map of the results of a corridor model for black bear movement between two disconnected portions of a National Forest.  The placement of the corridor was chosen in terms of an analysis of the elevation, land cover, and distance from roads- with its final size and position determined by the preferred habitat of the black bear.  Each cost, or factor affecting the bears' possible movement, was given a weight, and the GIS output was a predictive corridor between the two protected areas.  Possible movement of something as unpredictable as a wild animal can never be definitively plotted, but expanding a possible path into a corridor, with placement based on measurable factors, can provide a useful predictive tool.

Python Fundamentals, part I

This week we begin with the fundamentals of Python- how it is structured, and how it handles various types of data, like text, numbers, and lists.  Programming language, like communicative language, uses syntax, and the way various elements included are ordered and input makes the difference between a desired output and an angry red error message.

The output pictured above- my last name and the number 36- doesn't seem like much to produce, but the script used to create it took no small amount of effort.  It was not a simple "print" (to the screen) command to return those values, but rather a structured script that took my full name, put my first, middle and last names into a list, extracted my last name from the list, took the number of letters in my last name, and finally output that number times 3.  This process is completed in with Python script, using variable assignments, functions, methods and expressions.  It is a fairly basic procedure, and may seem a bit arbitrary, but the concepts learned and mastered therein are essential to the development of a decent working knowledge of Python.  It is upon these basic skills we build a greater understanding and mastery of the programming language.  

Land Use Suitability Modeling

If you've ever taken the time to consider, for some location, why something is there, whether it be man-mad or naturally occurring, you may be able to understand why land use suitability analyses and models are done.  The particulars of how these are typically structured and carried out is our topic for this week.

The above is a comparison of two methods of suitability analysis, both performed on raster data, and both created using a weighted overlay process.  The weighted overlay model assigns a weighted relative value to each raster cell, the difference between the two maps above being the equal weights scenario assigns the same level of importance to each of the five input values, and the alternative scenario assigns a different percent weight to each.  Each cell's value in the final output is a result of combining the relative values of the five inputs, which were raster data indicating certain criteria- such as land cover, slope, distance from rivers and roads and type of soil.  The values of the input cells for each criteria were based on a 1 - 5 scale, with higher values representing locations more desirable, such as those nearer to rivers and with certain types of soil.  By combining the five weighted values for the input (criteria) rasters, an output composed of cells indicating the relative suitability of an area, based upon the desired criteria, can be produced for a location like the study area in the maps above.  

An Introduction to Python & Computer Programming for GIS

A new semester, a new challenge- and this summer finds your humble blog author embarking on an endeavor to learn the basics of writing script for GIS with Python- ESRI's scripting language of choice for ArcGIS.  It is probably relevant to point out here that my prior experience with scripting in GIS, in its totality, involved listening to a particularly vocal co-worker complain about simple VBA operations for the version 9.2 ArcGIS we were using in those years, and I do recall him mentioning Python a few times as well.  Now (many years later) we are on ArcGIS version 10.2, and Python has been declared the victor in becoming the go-to script for automating various tasks in ArcGIS.

Python lists being open source and relatively simple/user-friendly among its advantages, and it is essential for any competent GIS professional to be at least familiar with its workings.  Even better, though, is to be at least somewhat well-versed in its use for automating geoprocessing tools, and that is the aim of this course.  We begin with learning the basics of opening the scripting and interaction windows in our Python editor- PythonWin.  From there we move on to becoming familiar with some of the basics on problem solving processes with scripts, and modeling with pseudocode and flowcharts.  Like any language, Python has its idiosyncratic ins and outs, and familiarity with some of their fundamentals is necessary as a start.

The above is a screenshot of a directory of folders that was created with a Python script.  The script was run, and the folders created, on the remote desktop that we use to access ArcGIS Desktop.  ESRI does, however, provide a free-of-charge, trial version of the software for students, good for one year.  Working on the remote desktop can be slow at times, especially on weekends when it seems that everyone in class is finishing their work, so it's often easier to use the software version run locally, without the remote connection.  The script provided for this week's exercise, that creates a directory of folders that will be used this semester, specifically references the drive used on the remote desktop connection though, and as-is must be run there to work.  Upon opening the script file in the PythonWin editor I found the specific line that does this with the simple command of "courseFolder = x:\xxx."  Running the script locally, on my laptop, I changed the location referenced (the "x"s in my example here) to the information for my personal hard drive, and ran the script there.  Lo and behold, it created the file directory (as in the screenshot above) right where I specified- on my computer.  This small triumph of successfully modifying provided code is to be savored, for the impression that these will be relatively rare this semester is rather striking.  Onward and upward we will continue though, on this exploration into the arcane and hallowed world of computer programming... 

Final Project: a GIS Assessment for the Bobwhite-Manatee Transmission Line

http://students.uwf.edu/emv5/IntroGIS/4043FinalProjectPresWeblink.ppsx

http://students.uwf.edu/emv5/IntroGIS/4043FinalProjectPresentText.pdf



The above are links to my final project- an assessment of the ways in which GIS can be used in planning a route through Manatee and Sarasota counties for FPL's new Bobwhite-Manatee Transmission Line, and what effects the line's route may have on local residents and the environment.  Included also is an estimate of the line's length, and its potential cost.

GIS is an integral tool for planning construction projects like this.  A 25 mile 230kV transmission line running through 2 very densely populated counties in south Florida is a huge project to undertake.  The route must be strategically planned, so that it doesn't effect too many local residents, and also shouldn't put too much stress on potentially fragile environments.  The potential route also must be accurately measured, so that an estimate of cost can be made.

All of the above objectives can only be completed with the assistance of GIS.  Maps can be used as powerful visual aids to convey some concern or idea, and can also be used to perform various analyses essential to the transmission line route's planning and construction.  

The Ultimate Thematic Map -or- "Now I actually know something about proper cartography"

At last the semester's end has arrived, and all of the skills and knowledge are put to the test: The Final Project.  It looms as a daunting specter, but its creation was nothing to dread, and I rather enjoyed the challenge. The task at hand was the creation of a thematic map depicting two different data sets- mean composite SAT scores for each state and each state's percent student SAT participation.

This is my cartographic opus of the semester.  I chose to display the percent participation as a choropleth map with six classes of data, and with the data classed using the natural breaks method, which necessitated some manual adjustment.  The natural breaks classes gave the best overall view of the trend in the data; it classed values with the most similarities together, and gave the greatest difference between class members and members of other classes.  The highest value class was manually created though, as those three scores were disparate enough from the rest to warrant it.  The mean composite scores (addition of the means in Critical Reading, Math and Writing categories for each state added together to form one average score) are depicted as graduated symbols, with classification using the natural breaks method, again using manual adjustments where outliers necessitated.  

The map itself is purple for the choropleth and yellow for the graduated symbols, in order to provide the greatest amount of contrast between the two.  Gestalt principles of visual perception dictate that the thematic data stands out best when like variables are symbolized with like variables, but the like variables are different in some key respect to allow for comparison of some measure.  This means that the student participation percentage for each state is symbolized with the state's shade of purple, so that it is immediately evident that the color of the state means the same measured variable across the map, but the different shades represent different classes.  The graduated symbols follow the same principle- they are all yellow circles, but their size varies according to the class of values they represent.  This representation for the display allows for easy visual identification of the two categories of information the map is displaying- each state's mean score and percent student participation.  The colors purple and yellow are ideal for this display, as they stand out well against one another, and help increase the visual contrast of the two measures.

The interesting thing here is what happens when the above tasks are completed and the picture of the data as a whole emerges.  One can easily see that the trend is for mean composite scores to decrease with increase in student participation, and increase as participation decreases.  If the mean scores and participation percents are plotted on a graph the trend becomes very evident- which is why I included the graph at the bottom of the map.  This obviously doesn't establish a causal relationship, but surely brings up some interesting and important questions.

I remain completely astounded about how little I knew about proper methods of map-making prior to this course.  I have been employed to make maps with GIS in a few different professional positions, and am now slightly mortified to think of my outputs therein.  The upshot, though, is that I feel as though the work I've done this semester has not only been extremely personally gratifying, but also a boon to my abilities as a GIS professional.  I aim to continue my career in GIS, and can absolutely see where I will use the techniques and skills I've learned here on a regular basis.   

Georeferencing and Editing Data

The term "georeferencing" may not mean much to most people, but it's a pretty important concept for those of us working with GIS.  Aerial photos, images of the earth's surface taken remotely, can really only be of use in GIS if they are an accurate spatial representation- otherwise they're really just photos.  These raster images can be lined up with an accurate map using control points, and linking places on the aerial photo with places on the map, and thereby stretching and warping the raster graphic as necessary to be spatially correct. 

The aerial images on the left side, the north and south portions of the UWF campus, were georeferenced to a vector base map of campus buildings and roads.  The text in the upper left corner of the northern photo, and the lower left corner of the south, details the root mean square error of the polynomial transformation.  This measure can be lowered by manipulating the control points used to align the aerial photo with the accurate map.  Control points create a more accurate transformation when they are placed as far from one another, and in as many different portions of the map as possible.  The 2 labeled features- the campus gym and Campus Lane- are edits, added to their respective vector dataset feature classes by the addition of a polygon and a line.  These differ from a simple polygon and line drawn on the map in that they are spatially referenced, and their shapes and attributes are actually added to the datasets permanently.  The eagle nest location on the right has 2 buffer zones added around it, which represent a conservation easement around the point feature.  

Google Earth & KML

GIS is a powerful tool, not only for making maps that previously would have been painstakingly time-consuming to create by hand, but also for easy and rapid spatial analyses.  This week we explored some of the ways that modern GIS and cartography are changing with new technology and the modern zeitgeist, and delved into creating KML files for use in Google Earth.

This screenshot is a Google Earth image of downtown St. Petersburg, taken from a KML file with a "tour" of various locations in south Florida.  The map layer information, visible when the layers are turned on and the image is zoomed out, was created by exporting the info directly from ArcGIS to the KML file format, and loaded into Google Earth.  Though the Google platform lacks the ability for visual modification and analysis present in ArcGIS, it is a free download, as opposed to the many hundreds of dollars required for the ESRI product.  KML files create an opportunity for sharing GIS information that previously was only available to those with the necessary software, as anyone with a computer and an internet connection can download and use Google Earth free of charge.  

Mapping in 3 Dimensions

One of the more interesting developments of creating maps in the 21st century is the ability to to create and view maps in 3D.  A 3 dimensional map can be a superior visual method to portray features both above and below the earth's surface, and can also assist in spatial analysis.  

This is a jpeg export of a 3D scene of Crater Lake, Oregon, to which I have added vertical exaggeration for effect.  One of the more creative aspects of 3D mapping is the cartographer's license to modify features as such- as a tool to better illustrate the map's vertical relief.  When opened in ArcScene the above graphic can be navigated in 3 dimensions, with the surface rotated to show the relief above, or below.  Different layers can also be added to this surface for spatial and/or visual analysis.  There are increasing opportunities for these kinds of 3 dimensional portrayals in GIS and map-making, which will likely only increase with future advances in computers and technology.

Geocoding and Network Analysis (and bus maps I never figured out how to make)

I, like most people, have a GPS application on my phone, and when I need to find directions I put in the address and go.  Google Maps does the heavy lifting for me.  What exactly goes into finding an address though?  The process, in reference to GIS, is called geocoding, and it is used in conjunction with network analysis to plan optimal routes for applications in things like EMS and public transportation.

Geocoding addresses is tantamount to querying a database and spatially locating locations associated with house/building numbers, streets and zip codes.  The address locator itself is quite customizable though, and can be modified according to the format of the input addresses.  The Network Analyst extension in ArcGIS can then be used to create a map like the one above on the left- which depicts a route from an address (11 Yorkshire Drive) to another location.  The useful thing about the brown line and the numbered points in that map is that they are more than just graphics of a line and some dots- they are actually recognized by the program as an interconnected route and stops.  Several years ago I had a GIS position with county transit, and created bus maps.  I remember drawing lines and points on a map as graphics, and cursing my inability to get the lines and points to "recognize" one another as stops and routes.  Clearly this would have been the ideal solution for that problem.