Dubai is one of the favorite tourist destinations, and hence, is the center for a wide array of events. Often, there is a shortage of taxis at such events, which can be a major issue in terms of transport. I have developed a software solution to optimize the deployments of vacant taxis to such events. 
My solution incorporates Machine Learning to predict how many taxis an event would require (say 'n' taxis) based on the event type, location, time of the event and estimated number of people attending the event. These attributes of the events can be mined from online websites and sources such as eventbrite.com or meetup.com. 
I have used Google Maps’ APIs to mark the event location on a map and finds ‘n’ number of required taxis which would reach the event location in the shortest time.
All of this is integrated into an app that is meant to be maintained by the Roads and Transport Authority (RTA) of Dubai. It should dynamically pull out information about events in real time from the web and notify taxis to proceed to these events. 
However, for my prototype, the events are supposed to be manually entered into a GUI, and the designated taxis for the event are displayed on a map.