Determination of development status of transportation facilities in Turkey on the basis of cities by using clustering analysis

The concept of transportation is one of the most important concepts that directly affects vital issues that are indispensable in a civilization such as education, health, economy, defense by providing connection and relocation between people from past to present. For this reason, developed states have always been in a more advantageous position throughout history. It is no coincidence that the transport facilities and infrastructure of the most prosperous societies are strong. This development process in transportation, which started with road transportation in human history, has increasingly diversified in different branches with seaways, railways, pipelines and airlines until present times. Forming the transportation activities between 2004-2018 year, 40 transportation indicator of Turkey in this study was made considering 81 provinces to analyze the situation in terms of transportation basis and will follow clustering analysis methods to help the transport policy are used. Here, it is aimed to draw an inference about the similarity of provinces relative to each other in terms of transport and the development status of provinces. Clustering analysis is a multivariate statistical technique that aims to divide a group of observations into a set number of clusters that are similar (homogeneous) and not similar (heterogeneous) between them and study them by various methods. Min-max standardization and Euclidean distance measure were used with R 3.5.1 and SPSS 16.0 programs in the data set. Among the cluster analysis methods, six hierarchical clustering methods (single linkage, complete linkage, average linkage, ward method, median linkage, centroid method) and two non-hierarchical clustering methods (k-means, k-medoid) were used. In addition, 14 cluster validity indexes were used, with the lower limit 5 and the upper limit 12 sets, to determine the number of clusters and the most appropriate cluster number was decided to be 5. As a result of the applied analyzes, Istanbul came to the fore with a big difference in the form of a single outlier in all cluster analysis methods. Among the hierarchical methods, the method that shows the best value of the cophenetic correlation coefficient used to measure the success of cluster analysis methods was the single connection method. According to the result of this method, Istanbul, Izmir, Ankara and Antalya were separated from other provinces. In non-hierarchical methods, Marmara, Aegean Region and Mediterranean Region came to the fore as provincial groups, and Eastern Black Sea, many Eastern Anatolia and Southeast Anatolian provinces emerged closely. Samsun and Trabzon provinces looks some what different in the Black Sea Region. The results of the analysis indicate that metropolitan provinces (provinces with a large population) are different in terms of transportation compared to the remaining provinces.