Initial results in dimensionality reduction of taxi DropOut-PickUp regions

Initial results with respect to dimensionality reduction of taxi PickUp-DropOut regions from New York City, Manhattan region, YellowCab company (2018 year, first 7 months). The dimensionality reduction is done separately for all working days and weekends using t-SNE, an SVD, and a simple deep autoencoder. The clustering quality assessment in two-dimensional space in which dimensionality reduction is done is conducted by using Silhouette, Calinski-Harabasz, and Davies-Bouldin metrics. Furthermore, the 15-minute taxi data aggregation is used.