Introduction
Clustering is a popular data mining technique useable in determining the groups of similar objects in a dataset. Clustering is a critical technique for gaining insights into a dataset’s structure and identifying relationships between entities (Weissman et al., 2020). The report will evaluate the clusters and provide descriptive, meaningful names for each collection. The information uses the method of example 17.4 to figure the clusters of cities based on the dataset in file p17_21.xlsx. The data in this file contains the population size, median income, cost of living index, and average housing cost of 32 cities. The research uses the k-means clustering method to divide these cities into four clusters.
The clustering of the cities using the variables makes sense. The clusters provide a way to group cities by focusing on the population size, income levels, cost of living, housing costs, crime rate, and unemployment rate. Each cluster is distinct from the others, suggesting meaningful differences between the cities in each group. The k-means clustering method begins by randomly assigning each city to one of four clusters. Then, the centroid of each cluster is calculated after identifying the strategy to use in the identification process. This centroid is the average of all the values of the variables for all the cities in the cluster. The descriptive, meaningful names for the collections provided above generate a way to identify and discuss the differences between the towns easily (Ambühl et al., 2021). Next, each city is reassigned to the cluster whose centroid is closest to the city’s values. The centroids are then recalculated for each set, and the process is repeated until the centroids no longer change.
Large, High-Income Cities
Cluster 1 cities are large, with a population of over 500,000 people. They also have higher incomes, with median household incomes above $50,000. The cost of living in these cities is relatively high, and housing costs are also higher than average. Crime rates in these cities are generally low, and unemployment rates are standard. These cities are well-off and have a higher quality of life than other cities in the dataset.
Medium-Sized, Lower-Income Cities
The cities in Cluster 2 are medium-sized, with populations of between 200,000 and 500,000 people. They have lower incomes, with median household incomes below $50,000. The cost of living in these cities is lower than in Cluster 1, and housing costs are also lower than average. Crime rates in these cities are generally higher than in Cluster 1, and unemployment rates are also higher. These cities are less well-off than those in Cluster 1, but they are still relatively stable.
Small, Moderate-Income Cities
Cluster 3 cities are small, with populations of less than 200,000 people. They have moderate incomes, with median household incomes between $35,000 and $50,000. The cost of living in these cities is lower than in the other clusters, and housing costs are also lower than average. Crime rates in these cities are generally moderate, and unemployment rates are relatively low. These cities are less well-off than those in Cluster 1, but they provide an affordable option for those looking for a place to live.
Small, Lower-Income Cities
On the other hand, cities in Cluster 4 are small, with populations of less than 200,000 people. They have lower incomes, with median household incomes below $35,000. The cost of living in these cities is lower than in the other clusters, and housing costs are also lower than average. Crime rates in these cities are generally higher than in the other groups, and unemployment rates are also higher. These cities are the least well-off of the four clusters and are likely to be more vulnerable to economic downturns.
Four unique clusters of cities were discovered after applying the k-means clustering method to the data. With comparable population numbers, median incomes, cost of living indices, and average housing expenses, Phoenix, Tucson, and Los Angeles were included in the first cluster, dubbed the more significant costs cluster. San Francisco, San Diego, and Sacramento were part of the second cluster, known as the expensive cities cluster. More significant cities had greater populations, median incomes, cost of living indices, and average housing expenses with comparable population numbers and median incomes but lower cost of living indices and average housing expenses. The third cluster of lower-cost medium-priced cities is Denver, Colorado Springs, and Albuquerque. They include the fourth cluster and Las Vegas, Reno, and Salt Lake City, with smaller population sizes, lower median incomes, lower cost of living indexes, and lower average housing costs.
Conclusion
The four clusters provide a way to group cities based on population size, income levels, cost of living, housing costs, crime rate, and unemployment rate. Each cluster is distinct from the others, suggesting meaningful differences between the cities in each cluster. The clusters are based on the similarities in population size, median income, cost of living index, and average housing cost among the cities. The clusters also make intuitive sense, as the larger cities with higher prices have been grouped, as have the smaller cities with lower costs. The descriptive, meaningful names provided for the clusters also accurately describe the cities within each cluster.
References
Ambühl, L., Loder, A., Leclercq, L., & Menendez, M. (2021). Disentangling the city traffic rhythms: A longitudinal analysis of MFD patterns over a year. Transportation Research Part C: Emerging Technologies, 126, 103065. Web.
Karakoyun, M. (2019). A new approach based on K-means clustering and shuffled frog leaping algorithm to solve travelling salesman problem. Academic Perspective Procedia, 2(3), 446-453. Web.
Weissman, B., van de Laar, E., Weissman, B., & van de Laar, E. (2020). What are big data clusters? SQL Server Big Data Clusters: Data Virtualization, Data Lake, and AI Platform, 1-10. Web.