benchmarks

Currently, SZTS includes five representative benchmarks, stzod, hotregion and mapmatching are implemented by java, hotspot and secsort are implemented by Apache pig.

sztod

The sztod program computes how many moving objects start from point A and head for point B in a given time interval in the city

hotregion

The hotregion program computes the distribution of people, cars, or other vehicles in Shenzhen within a given time interval.

mapmatching

mapmatching program is used to match the observed GPS trajectory with the route on a digital map.

hotspot

hotspot analyzes the intensity of traffic flow (free or congested) to identify the traffic phases of each hotspots. In addition, hotspot contains a prediction model based on logistic regression to predict the future traffic flow intensities for a given hotspot.

secsort

In a metro application, secsort first sorts the records by smart-card record, then sorts by the primary time-stamp key, and finally sorts the records in each group by the secondary metro station.

data source

There are three primary data sources in the Smart Urban Transportation System of Shenzhen.

(1)public bus transportation system

It produce GPS record data and smart card transaction data

(2)taxi transportation system

It produce GPS record data and deal transaction data

(3)subway transportation system

It produce smart card transaction data

data format

each record has many fields, each field is described in szts documentation.

data scale

Shenzhen is a famous international city located in southern China. It covers 2000 square kilometers with a population of approximately 18 million. Shenzhen already built a modern urban transportation infrastructure including 5 subway lines with 118 stations, 936 bus lines, and 30,000 cabs. Each bus and cab is equipped with a GPS device. More detail information about SUTS are listed in table.