Assignment 1

Goals and Background

The following are the goals of our first assignment:

• Differentiate between levels of measurement
• Differentiate between classification methods
• Retrieving data from the U.S. Census and Joining Data
• Enhance Cartographic Knowledge

Part 1

During the first part of the assignment I will be defining (in my own words) the difference between Nominal, Ordinal, Interval, and Ratio data. Additionally, I will be providing examples in the form of maps for each of the types of data.

Nominal

Nominal data is essentially a specific category which the data is classified under. Nominal data has a title or name which it is classified under. Examples of nominal data are Land Use/Land Cover (LULC) type (Fig. 1), breed of cattle, or sex (male, female). Examining Fig. 1 you can see the map is broken into categories of LULC. The categories associated with type of LULC are examples of nominal data. Additionally, the highlighted counties (Trempealeau and Buffalo) are another example of ordinal data. 

(Fig. 1) Map I created using nominal data for a previous class project.

Ordinal

Ordinal data is comprised of data which can be rank compared to other forms of the same data. Examples of the ordinal data would be a ranking states in order of their establishment (Fig. 2). The numerical value given in ordinal data doesn't tell you specifics about the variation other then which one is first, second, ect. Example in Fig. 2, you cannot tell how many years later California was established after Wisconsin.

(Fig. 2) Map of states in the ranked in order which they were established. Borrowed from: https://www.printableworldmap.net/preview/Statehood_Order_Map

Interval

Interval data unlike ordinal data the difference between the given values have meaning. However, in interval data the "zero" doesn't determine the bottom of the data. With the "zero" just being a number on the scale one cannot create a ratio interpretation from the data. One of the most common examples of interval data is temperature on the Fahrenheit scale.

(Fig. 3) Temperature map from a weather station which is a great example of interval data. Borrowed from:http://www.usairnet.com/weather/maps/current/current-temperature/

Ratio

Ratio data is more common than interval data. Ratio data unlike interval data does have a "zero" which is the bottom or the start of the data. The zero allows you to create a ratio between the numbers. Kelvin temperature scale is an great example of ratio data. Kelvin temperature start a 0 and only goes up. From ratio data you are able to create an true "ratio". Example 4 degrees Kelvin is twice as warm as 2 degrees Kelvin. Height and weight of a human is also another example of ratio data. 

(Fig. 4) Ratio map of boys compared to girls in China using ratio data. Borrowed from:https://sites.google.com/site/onechildpolicysuth/real-world-data/charts-and-other-data

Part 2

I was given the following scenario for part 2 of the assignment:

"You have recently been hired by an agriculture consulting/marketing company and have been asked by your boss to provide a number of maps to be presented to potential clients.  Specifically, your company is interested in increasing the number of women as the principal operator of a farm.  Where do you think your company should concentrate the message to increase farming among females?" 

I am to create three maps using Equal Interval based on Range, Quantile, and Natural Breaks classification methods. I will utilize the same data and display it at the county level for all three methods. After the creation of the maps I am required to create a report defining the methods used for each map and present the maps. Finally, I will choose the map which I feels best represent the method for the given assignment and provide a convincing explanation to my reasoning. My professor provided me the data which he obtained from the United States Department of Agriculture: Census of Agriculture 2012.

Before heading to the results I will provide you with a brief definition and description of the classification methods I will be using. Additionaly, I will provide an image of the histogram in ArcMap with the class breaks (blue lines) displayed for each method (Fig.5-7).

Range

Range classification breaks the data into a "range" such as 1-10, 11-20, 21-30, 31-40. The range method I selected in ESRI ArcMap was Equal Interval. The Equal Interval setting breaks the data into an interval range based on the data range. I chose this method to best show the variation which can be had through various classification methods. Side note for the data I was provided is the data range was not a whole number but I changed the Symbology labels for the legend to show whole number, as you cannot have .25 of a person.

(Fig. 5) Range (Equal Interval) classification histogram breakdown in ArcMap.

Quantile

Quantile classification breaks the data into classes of equal numerical values. If you choose to group your data into 4 classes and you have 8 data records the Quantile classification will put 2 data records into each class.

(Fig. 6) Quantile classification histogram breakdown in ArcMap.

Natural Breaks (Jenks)

Natural breaks classification is the toughest classification method to describe. Below is the ESRI definition:

A method of manual data classification that seeks to partition data into classes based on natural groups in the data distribution. Natural breaks occur in the histogram at the low points of valleys. Breaks are assigned in the order of the size of the valleys, with the largest valley being assigned the first natural break. (support.esri.com)

I define Natural Breaks as the clumping of data in to groups based on the clustering of the numerical values of the data. Example: A 10 number data set has the following: 1,3,3,5,5,6,8,8,9,9. The directions were to group the numbers into three groups which would result in the following groups if I was preforming the classication:

1. Group 1: 1,3,3
2. Group 2: 5,5,6
3. Group 3: 8,8,9,9 

Natural breaks classification tends to be my go to method for the majority of geography depending on the application and the specific data set. This is not to say it is the only way but I feel it provides the best representation. See the Discussion section for my review of these three methods.

(Fig. 7) Natural Breaks classification histogram breakdown in ArcMap.

Results

Range Map

Quantile Map

Natural Breaks Map

Discussion

Analyzing the results, my first determination is not to use the Quantile map. The quantile classification method grouped the values in to ranges which are misleading to the reader. Specifically, the class of 140-386 female farm operators is poorly grouped with a very broad span of numerical values compared to the 1-46 class.

As I stated before I am a fan of the Natural Breaks classification method. However, after examining the results I concluded the natural breaks methods gave a misleading result to the data. Similar to the Quantile results the 4th class (187-386) was to broad of a range compared to the other classes. The results from this method are better than the Quantile but still not the best display.

My favorite from these three maps is the Range (Equal Interval) classification. Breaking the classes into an even range provides the best display of the data in my opinion. Having equal ranges provides a good representation throughout the state without any misleading distortion.

With that being said I don't believe any of these maps should be used to represent female farm operators. More data needs to be provided to give a accurate representation of the operators. Information such as the number of farms or the acres of farm land per county would be very beneficial to normalize the data. The normalization of the data would provide a better representation of the female farm operators in the counties of Wisconsin.