To date, GPS tracking data for minibus taxis has only been captured at a sampling frequency of once per minute. This is the first GPS tracking data captured on a per-second (1 Hz) basis. Minibus taxi paratransit vehicles in South Africa are notorious for their aggressive driving behaviour characterised by rapid acceleration/deceleration events, which can have a large effect on vehicle energy consumption. Infrequent sampling cannot capture these micro-mobility patterns, thus missing out on their effect on vehicle energy consumption (kWh/km). We hypothesised that to construct high fidelity estimates of vehicle energy consumption, higher resolution data that captures several samples per movement would be needed. Estimating the energy consumption of an electric equivalent (EV) to an internal combustion engine (ICE) vehicle is requisite for stakeholders to plan an effective transition to an EV fleet. Energy consumption was calculated following the kinetic model outline in "The bumpy ride to electrification: High fidelity energy consumption estimates for minibus taxi paratransit vehicles in South Africa".

Six tracking devices were used to record GPS data to an SD card at a frequency of 1Hz. The six recording devices are based on the Arduino platform and powered from alkaline battery packs. The device can therefore operate independently of any other device during tests. The acquired data is separately processed after the completion of data recording. Data captured is initiated with the press of a button, and terminated once the vehicle reached the destination. Each recorded trip creates an isolated file. This allows for different routes to be separately investigated and compared to other recordings made on the same route.

There are 62 raw trip files, all found in the attached 'raw data' folder under the corresponding route and time of day in which they were captured. The raw data includes date, time, velocity, elevation, latitude, longitude, heading, number of satellites connected, and signal quality. Data was recorded on three routes, in both directions, for a total of six distinct routes. Each route had trips recorded in the morning (before 11:30AM) , afternoon (11:30AM-4PM) and evening (after 4PM).

The processed data is available in the 'Processed Data' folder. In addition to the raw data, these processed data files include the displacement between observations, calculated using Geopy's geodesic package, and the estimated energy provided by the vehicle's battery for propulsion, braking, and offload work. The python code for the kinetic model can be found in the attached GitHub link https://github.com/ChullEPG/Bumpy-Ride.

Future research can use this data to develop standard driving cycles for paratransit vehicles, and to improve the validity of micro-traffic simulators that are used to simulate per-second paratransit vehicle drive cycles between minutely waypoints.

To date, GPS tracking data for minibus taxis has only been captured at a sampling frequency of once per minute. This is the first GPS tracking data captured on a per-second (1 Hz) basis. Minibus taxi paratransit vehicles in South Africa are notorious for their aggressive driving behaviour characterised by rapid acceleration/deceleration events, which can have a large effect on vehicle energy consumption. Infrequent sampling cannot capture these micro-mobility patterns, thus missing out on their effect on vehicle energy consumption (kWh/km). We hypothesised that to construct high fidelity estimates of vehicle energy consumption, higher resolution data that captures several samples per movement would be needed. Estimating the energy consumption of an electric equivalent (EV) to an internal combustion engine (ICE) vehicle is requisite for stakeholders to plan an effective transition to an EV fleet. Energy consumption was calculated following the kinetic model outline in "The bumpy ride to electrification: High fidelity energy consumption estimates for minibus taxi paratransit vehicles in South Africa".

Six tracking devices were used to record GPS data to an SD card at a frequency of 1Hz. The six recording devices are based on the Arduino platform and powered from alkaline battery packs. The device can therefore operate independently of any other device during tests. The acquired data is separately processed after the completion of data recording. Data captured is initiated with the press of a button, and terminated once the vehicle reached the destination. Each recorded trip creates an isolated file. This allows for different routes to be separately investigated and compared to other recordings made on the same route.

There are 62 raw trip files, all found in the attached 'raw data' folder under the corresponding route and time of day in which they were captured. The raw data includes date, time, velocity, elevation, latitude, longitude, heading, number of satellites connected, and signal quality. Data was recorded on three routes, in both directions, for a total of six distinct routes. Each route had trips recorded in the morning (before 11:30AM) , afternoon (11:30AM-4PM) and evening (after 4PM).

The processed data is available in the 'Processed Data' folder. In addition to the raw data, these processed data files include the displacement between observations, calculated using Geopy's geodesic package, and the estimated energy provided by the vehicle's battery for propulsion, braking, and offload work. The python code for the kinetic model can be found in the attached GitHub link https://github.com/ChullEPG/Bumpy-Ride.

Future research can use this data to develop standard driving cycles for paratransit vehicles, and to improve the validity of micro-traffic simulators that are used to simulate per-second paratransit vehicle drive cycles between minutely waypoints.