To make our routing more efficient and easy to use, we can further improve our project in the following aspects. 1.Design additional sensors on trash bins to detect whether the trash there is stink. In our project, we tried to look for correlation between trash level of each bins and its nearby air quality, it seems that the street level AQI is not local enough to reflect the influence of full trash bins. Thus an special sensor mounted inside the trash bin may help significantly if we want to prioritized the collection of stink trash. 2.Machine learning methods may apply to the predict future trash level of each trash bins. Our project now has only plan the route based on the real time trash level, which is accurate and efficient but not easy to be planned ahead of time. Thus if we build a machine learning model to predict future trash level of each bins, we can plan a whole month’s collection route Therefore manpower and equipments can be scheduled and cost can be reduced. 3.We can build mobile-friendly client to improve user’s experience. In our project now, web application are implemented as only client to the users. Web application is universal and platform-independent but not efficient and optimized enough. Thus mobile friendly application can significantly improve the experience. For example, special iPad application can be very useful because it is suitable for trash truck’s working condition. 4.Data collection process can be further optimized. In our project now, we spider the data from BigBelly’s management website in a brute-force way. This is either not efficient nor elegant to use all the data. If later we can directly access the database of the management website, it will be much easier to utilize all the data and result a more functional project.