Prof. Ghaddar worked with colleagues at IBM and two universities in Italy to analyze the Twitter feeds of potential carpoolers to gain insights into their personal interests. They entered this information into a computer algorithm that matches carpoolers based on their locations, time preferences, and personalities. The results showed that if carpoolers were compatible, then people's satisfaction increased and car use decreased.
The research was published in Transportation Research Part C: Emerging Technologies, and was featured on CBC, The Record, Science Daily, and IEEE Spectrum, among others.
You can listen to an interview by Prof. Ghaddar for CBC Radio here.