Computing our way to electric commuting in Africa: The data roadblock

Abstract

In the information age, the value of good data can hardly be overstated. Unsurprisingly, a severe lack thereof is seen as a pothole on the road to the decarbonisation of Africa's paratransit -- the mainstay of the region's transport. Data acquisition in transport used to be concerned with the temporal flow rates and volumes of passengers and vehicles on sections of roads. This was done with surveys and vehicle counting. Separately, vehicle mobility data has been used for vehicle health monitoring, driver behaviour monitoring and vehicle recovery. For paratransit in Africa, where passenger counts and routes are unknown (and fluid), and where data acquisition is tedious, standardised passengers with tracking applications on phones are extensively used for this purpose. This brings the meaning of “good data” in this region of low and lower-middle-income countries into focus. Given the long ranges and fast refilling times of combustion engine vehicles, manufacturers and fuel outlets had hitherto existed in a symbiotic relationship without the bondage of mobility pattern information. However, in the era of electrification, battery-powered vehicles and their specifications and limitations have become inextricably coupled to road-side infrastructure through their mobility patterns due to their lower range and slower charge times. The infrastructure energy supply side can be modelled with dated passenger-based or roadside-based gathered data, assuming unchanged patterns. But, the charging potential calculations requires stationary rather than moving times. Moreover, energy demand per point is effectively that of an individual vehicle, which depends on spatio-temporal mobility patterns and battery capacity. Intermittent renewable energy generation further complicates this time variant challenge. In this paper we use a public passenger-based data set that is normally used to characterise routes, which has over 300,000 trips, to establish the energy requirements of seven cities in Africa - Abidjan, Accra, Cairo, Freetown, Harare, Kampala and Nairobi. The results show that the peak demand per city varies wildly, apparently with the reliability of the city's data, with realistic peaks of 100 MW to 300 MW. Although the results give an indication of the supply side requirements, it highlights the problem with using incomplete and/or unreliable data to estimate a city's peak load, which points to a need for vehicle-based data acquisition to adequately answer the question, or at least validate, the results.