by Bernhard Niedermayer
For the DMA Mobility Pilots, the first months of the DMA project were focused on requirements definition and data acquisition. The objective is to develop two pilots in the mobility domain as showcases on how the DMA platform can be used to created novel data products and services based upon existing data.
Demands prediction
The use case of the first pilot is to predict the future demand for taxi rides throughout a city to optimize entire taxi fleets. The formal requirements specification is currently about to be finalized. Nevertheless, the most important design parameters have been agreed upon already. We will focus on the Vienna area, divide this area into cells of a uniform grid, and predict the level of demand for each cell.
Current hypothesis are that the likelihood of passengers choosing a taxi over other means of transport depends upon several static influences, such as the season, the day of week, or the time of day – but also upon dynamic factors, such as current and near-future weather conditions or time pressure (as caused by delays of intercity rail connections, for example).
In additon, to this likelihood for choosing a taxi, the number of persons present at a certain location determines the predicted demand. To this end, several data products have made available by now, such as weather data, mobility data, train delays, or public events. Also, information about actual taxi rides has been contributed by Taxi 40100. This will allow us to verify our hypothesis.
A first, rudimentary version of the heatmap service has been implemented – relying on train delay and event data which is publically available. Next steps will focus on integrating more data sources into the heatmap generation as well as verifying our assumptions about actual the demand for taxi rides.
Future pilot in the pipeline
The second pilot “Mobility-based Energy Prediction” is still in the requirements engineering phase and will be presented in detail soon.