Richardson – Week 2

Eliza Richardson 

27 January, 2023 

ArcGIS Week 2 Blog Post

Chapter 1 

Mitchell begins the first chapter with laying out some necessary information needed for GIS mapping, and how to decide the best way to represent your chosen data. You first need to understand the data, choose a method, process the data, and then contribute results. These results could either be discrete (location of businesses in the state of Ohio) or continuous (counties color coded by number of businesses). They would both display the same data, but with a different application and meaning of the data. Discrete data can be represented with a vector model, which is better for individual points, and continuous data can be represented with the raster model, which is better for widespread data. Mitchell then describes the types of attributes that you can describe your data with; categories, ranks, counts, amounts, or ratios. An example of categories could be types of employment in Franklin county. You could categorize each household on what kind of field they are employed in. Ranks may relate to things like the redlining maps; which neighborhoods are more likely to be accountable for a loan and pay a mortgage on time. Then be sorted into ranks of A – D. Counts and amounts refers to any measurable quantity, such as the amount of employees at various businesses. Ratios can be separated into proportions and densities. For example, you can map the average number of people per household by county to see which counties have the most people living in each home. 

A lot of the information covered in this chapter are things that I have heard discussed and have used when doing GIS projects in other courses, but it is very helpful seeing them all spelled out in front of you so that you can maximize the efficiency of your data. In other courses, I felt like I wasn’t presenting my data in the most optimal scenario, and I could have used a better consensus of the way to maximize the presentation of it. 

Chapter 2

GIS is so important in the development of so many different fields. Using GIS, “police can map where crimes occur each month, and whether similar crimes occur in the same place or move to other parts of the city.” and “wildlife biologists studying the behavior of bears may want to find areas relatively free of roads to minimize the influence of human activity.” (Mitchell 24) In order to create a map that represents your data, you first need to link your data to geographical coordinates. You can then link the subset of your desired data to the geographical coordinates in order to create the image you want to portray.  However, if you don’t want to portray too much information to digest in your figure, then it will be too difficult for the reader to understand the purpose of the map. If you have a detailed description, you could categorize them into smaller portions in order to make the message of the figure clearer. Mitchell states that no more than 7 categories should be used on a map or else the message of the data will be lost. The fewer categories you can evaluate in one image, the easier it makes for the reader to understand and compare each subset. In addition, the scale of the map and the amount of categories can make the patterns difficult to see for the reader, and can get lost in the figure. Keep It Simple Stupid. 

However, if you need to include all of the categories in order to portray the message of the data  completely, one way to do this is to create multiple maps of the same geographical area, but with different data that it is showing. If you want to show 15 data points, you can subset them into categories, and then make another map for each category that you subset the data into. This makes it so that you can still provide all of the data necessary to reach a full conclusion, but still allow for the maps to be decipherable. 

Chapter 3

When exploring data, it is important to keep in mind the purpose of the figure, and what the figure should be telling the reader. Are you evaluating the data? Or are you trying to find a pattern? Or an answer? Based on the goal of the data, then you can choose the way you want to graph your data. One way that you can evaluate one data point is through classes. This is very similar to rankings, but you can assign an upper and lower limit to which values fall within the rank, to represent the entirety of the region. For example, if you are trying to evaluate soil quality, you can assign a value of 8 to the top 15% of soil, and separate all soil regions based on the following 15% to show which areas have the best soil quality. You could also classify a region into percentiles, and represent the bottom 25th percentile, middle 25-75 percentile, and top 75th percentile separately. Mitchell goes through the various ways that you can section data. Keep in mind that not all of these sectioning strategies will be optimal with your data, and you must choose depending on the layout of your data.  Natural breaks (jenks) which is when data is separated based on where there is a jump in values. Natural breaks are good for mapping a data set that is not evenly distributed and has clusters of information. Quantile contains an equal number of features. Quantile is good for comparing areas that are roughly the same size, and for data points that are relatively equally distributed. Equal interval is when the difference between the high number and the low number are the same for each section. Is good for data that does not have a large variance in value and for presenting nontechnical information like precipitation and temperature. Standard deviation is when “features are placed in classes based on how much their values vary from the mean.” (Mitchell 68) This is good for seeing which values are above and below the mean, and how far above and below they are. 

Chapter 4

Mapping density is very important when trying to distinguish trends in the area. There are two ways to create a map by density: “based on features summarized by a defined area, or by creating a density surface.” (Mitchell 109) By using a defined area to create a density map, you can use predetermined boundaries, such as counties or countries, to determine the density of that area as a whole, compared to another section of the region. By using a density surface, you can see the variation in density across the area as a whole, not simply where boundary lines occur. You should use a density map if  “you have data already summarized by the area, or lines or points you can summarize by the area”. This is easier than creating a density surface, but can cause some inaccuracy, especially if you are analyzing a large area. You should use a density surface if “you have individual locations, sample points, or lines.” This requires more data processing on the authors part, but is more precise in the long run. Another way you can add to a density map is to add a layer of the density points overtop of the map. That way you can see where the individual points are coming from, but also the overall trend for boundaries in the data. Another way to change the variability of the density map is to change the cell size. If it is hard to distinguish where the dense places are on the map, try making the cell sizes larger so that you get a closer look at parts of the data. To distinguish non dense areas between dense areas, you should use a color gradient to make a visual representation of how the density fades between geographical regions. However, if you have too many gradient points, it will be difficult to distinguish where the points are falling on the graph, and the difference between each region. Again, you don’t want to include too much information so that it is hard for the reader to comprehend.