Along with the rapid expansion of digital music formats, searching expected songs from Internet has become easy. However, for some specific conditions such as driving a car, manipulation of managing track playlists may cause danger. From this perspective, we build iCarRadio, an in-car music recommendation system based on the user’s surrounding conditions. We collect luminosity from a light sensor, location and speed of vehicle from a GPS, weather and traffic information from a REST API, and then gathered them in Intel Edison. Tracks will be recommended with the help of Amazon cloud services and Spotify API based on the real-time information.
In our project, we use a light sensor, a GPS, a weather API and a traffic API to collect data. The gathered data will then be sent to Amazon Web Service by Intel Edison, and data processing will be implemented by Amazon cloud services. Machine learning prediction results, in our case which are audio features, are calculated by AWS and then sent to Spotify API. A list of best-matching songs will be added to the user’s playlist.