VIZUALIZING THE POTENTIAL OF A PREDICTIVE IN-CAR RECOMMENDATION SYSTEM
The syn(c)ity video illustrates the potential of a predictive, in-car recommendation system based on the real-time profiling data on the city and the driver. Key to the system are novel methods of filtering relevant data to provide to any given driver in any given time and place.
In the video, different typologies of information are associated with different visual elements that form a structured overall grammar.After introducing the various real data sources that were used to generate the visualization, emphasize is put on three distinct filtering approaches related to the driver. Considering the massive amounts of real-time data potentially available to a driver and considering the imperative of reducing distraction from the driver’s attention on traffic, filtering techniques that can extrapolate the few but very relevant data for a given driver in a given time and space are increasingly critical. The filtering modalities presented are thus:
Filter by distance: Data is mainly filtered by proximity to the vehicle position. The user can set the desired distance radius and additional filtering semantics can be added.
Filter by personal interests: Data from social network profiles as well as travel predictions as described above are used to infer relevant data and informations to be conveyed to the driver.
Filter by user query: Given the city and driver profiling of the system, a user’s query does not need to focus on indicating a place directly. The place is a result of the query in this case and subsequently additional real-time information is conveyed to the driver in function of reaching his overall goal expressed in the search query.