Data is consuming our lives. New methods for collecting and analyzing data are constantly being evolved and introduced. However, one of the most consistent and available sources of data has yet to be explored: trash. Americans on average consume around 90,000 lbs of trash in their lifetime. Canalytics explores this new field of data through long-term analysis on consumer's daily waste habits. Furthermore, it uses this data to attempt to reduce consumer carbon footprint through consumption suggestions and real-time shopping notifications.
What it does
Canalytics is a cloud-based object recognition system at its core. Consumers are too busy to input data regarding what they eat, and so we do it for them. Our hack consists of a Raspberry Pi-powered trash can that uses machine learning to identify garbage and classify it as recycling or trash. This data is then fed into a data analytics platform to determine long-term consumer trends and identify ways to improve consumption habits.
Challenges we ran into
Two facets of the project that were most troubling were the mechanical design and cloud integration. Throughout the night, we spent several hours testing different design models for our physical trash can and its motorized sorting. Additionally, we spent several hours trying to bring together several cloud-based computing applications in order to fully synchronize parts of our project.
Accomplishments that we're proud of
We are most proud of our group cohesion throughout the event. While many of our individual applications relied on the completion of another team member's work, we were able to delegate responsibilities effectively and work very efficiently to accomplish our goal.
What we learned
What's next for Canalytics
Depending on the reception of Canalytics, we are interested in further developing the consumer model of our product and refining how we can use the data generated from our IoT trash can. This further development would include additional product design, refinement of our machine learning algorithms, and the creation of our own data analytics API. We would also love to work with the University of Michigan to build upon our project in order to develope valuable insights about student waste on campus.