Spatial Patterns
In this section, we will see the distribution of pick up or drop off points at the airport. The heatmap and DBSCAN clustering results will help us see where clusters are located. The analysis is done in R and ArcGIS Insights.
Where are the points?
On the left is the distribution of points in January (containing 97,312 points). It looks somewhat dispersed, although there seems to be more points around the two terminals.
Heatmap
A heatmap is created to show the hot spots. Generally speaking, there are two most popular pick up/drop off places: around Terminal 1 and Terminal 2.
Clustering
Density-based Spatial Clustering is suitable for clustering of geographic location data. The first 3 clusters in the result contain the majority of points and are visualized on the left:
- Terminal 1 (red)
- Terminal 2 (green)
- Economy Shuttle Stop 9 (blue)
Distribution of time spent in airport of the 3 clusters
On average, vehicles spend 35.7 min in cluster 1 and 34.7 min in cluster 2. In cluster 3 drivers spent the least: about 6.9 min ("Time" refers to the time from entering the airport until leaving the airport fence).
In the distribution plots, more cars in cluster 1 and 2 spend 10-20 min while more cars in cluster 3 only spend 5-10 min.
The analysis above may help both airport management better understand the vehicle spatial patterns and drivers avoid traffic congestion.
For Insights page, please visit here. You are also welcome to visit my Github repository for detailed R analysis.