Friday, April 12, 2024

Computer Cartography - Module 4 - Data Classification

 This week's lab module for Computer Cartography 5007L focused on understanding and comparing four different data classification methods. We were tasked with reviewing and preparing Miami-Dade County Census tract data and presenting it in two separate map compilations using ArcGis Pro. The first map showed the percentage of senior citizens in each census tract, while the second map showed the normalized count of senior citizens per square mile. Both maps used four classification methods: equal interval, quantile, standard deviation, and natural breaks. For my maps, I used a blue continuous choropleth scheme with the exception of the standard deviation maps. I originally had my initial classes as white however I changed this after reading in our textbook Cartography that this is not recommended. I found the layout to be a little bit more challenging than I anticipated but it was a great learning experience. The four classification schemes we followed are described below:

Equal Interval - The Equal Interval classification method divides attribute values into equal-sized ranges. While these ranges might be evenly spaced the number of records in the range or category can differ. This method is great for emphasizing the number of attributes relative to one another however, your data will not be distributed properly. This can lead to many features being in one class while in another class there might be none.

QuantileThe Quantile classification method divides classes into evenly filled ranges. This method is ideal for ordinal data with a clear ordering of variables. This method can be misleading because interval sizes can vary and similar values can end up in different classes. Also, widely differing features can end up in the same class. This distortion can be minimized by increasing the number of classes.

Standard Deviation - The Standard Deviation method shows the variation of the attribute values from the mean. This splits the values into above and below the mean value. The class breaks are created with equal value ranges from the mean value. These might be unevenly distributed but are not skewed toward either end. Standard Deviation can be advantageous because it uses all possible information however it can be disadvantageous if your data is not normally distributed.

Natural BreaksThe Natural Breaks method divides data into naturally occurring “breaks” found within a data set. This method seeks to minimize class variance and maximize variance between classes. This results in classes that are usually different from one another. This method can be advantageous because it can create a more accurate or unique class scheme for each map. A few disadvantages are that it is not ideal for data that has low variance and due to the data-specific classifications produced by the method comparing multiple maps with differing data is not useful.

The map below is my final product for the normalized senior citizen population age 65+ per square mile in Miami-Dade County, Florida. I believe the normalized population count best represents the data set because it more accurately depicts the distribution allowing the map reader to see that there is a concentration of seniors in the northeast portion of the county. The maps based on a percentage above age 65 contain some potentially misleading outliers that could distract the reader. The population count by normalized data does a better job at eliminating some of the outliers seen in the percent above 65 maps which leads to a better overall map that is easier to interpret especially for a novice map reader. I think the quantile method does a great job of presenting the data showing a clearer distribution than natural breaks which has an initial class that goes up to 864 in comparison to the quantile which only goes up to 440. For this reason, the quantile provides a better idea of the concentration of seniors and would likely be easier for novice map readers to interpret.

Overall, I found this lab to be very helpful in beginning to understand data classifications as they pertain to cartography. I have very little experience in this topic and hope to be able to continue to understand and grasp the concepts as I move forward.





No comments:

Post a Comment

Applications in GIS - Module 6 - Suitability & Least Cost Analysis

In Module 6, we learned about Suitability and Least Cost Path Analysis. We were introduced to performing suitability analysis using both vec...