At Rubbish, making our planet cleaner and greener is of utmost importance, and data plays a pivotal role in achieving this goal. Our latest initiative leverages the power of AI-enabled environmental mapping. Environmental data is collected through crowd-sourcing on the Rubbish app, training government agencies to collect data, or collecting the data ourselves.
But what exactly is the point of collecting data on trash and litter in communities? Proper management of waste and adequate city management requires high-quality data. Rubbish, with the help of AI integration, is able to perform environmental mapping of data at a great scale.
Rubbish not only deals with litter and pollution, but graffiti and infrastructure issues as well. For instance, users can report if a trash can is broken, contaminated, or in need of a deep-cleaning.
Users simply take a photo of the issue on the Rubbish app, and the app geotags the location, creating a data point on our environmental map. This provides a real-time, comprehensive view of litter hotspots and pollution patterns across different regions.
The Rubbish app’s AI integration categorizes the types of litter, which is a crucial time-saving advancement, as users do not have to manually identify types of litter themselves. Rubbish put this tool to use in the SOMA West community of San Francisco. In 2019, Rubbish tracked issues of concern to the community, prior to and after the launch of the SOMA West Community Benefit District. The massive improvements over time are shown below, and illustrate a use-case for Rubbish’s environmental mapping initiatives.
In another case, Rubbish compared key problems within the River District of Sacramento to the Downtown area. Surveying the different kinds of litter and other concerns within the communities, it became clear that vast discrepancies existed in these two neighborhoods, with more resources being needed in the River District. The results spoke for themselves, such that Jenna Abbot, Executive Director of The River Community Improvement District, stated, “How did my colleagues react? They were shocked, dismayed, extremely interested, and very motivated.” Environmental data mapping, in this instance, can be used to compare regions and the varying needs of each.
Environmental data collection of this kind leads to more effective management of communities. Local governments, environmental organizations, and communities themselves, can use this information to allocate resources where they are needed most. Funding for cleanup efforts and litter reduction programs is limited, so ensuring it reaches the right areas is essential.
Individual Rubbish runs, as shown above, can utilize AI categorization to map out litter locations. While these maps might seem small, when put together, user activities on Rubbish can lead to large, expansive maps that are very useful and data-rich.
The Rubbish app’s AI detection feature is also effective on heavily degraded items, shown in the image below. It accurately categorized textiles and glass litter, even though these items were covered in dirt and were highly damaged.
Incorporating more functionality is a priority, and Rubbish is improving brand recognition capabilities, as this video indicates.
Rubbish is always working on improving the app to make it more useful and accurate, especially in regards to environmental data collection. The Rubbish team is working toward integrating multiple-item detection, meaning that the AI assistant will be able to categorize more than one item from a single picture!
Accuracy is also being improved; our current model achieves an accuracy of greater than 85%. Rubbish is aiming for a +90% accuracy rate from the AI categorization feature by the end of this year.
Overall it is clear that environmental data is indispensable. Rubbish, through multiple methods of data collection, is able to map this data such that it can be utilized and viewed by municipalities, agencies, and citizens alike.
Download the Rubbish app here, and give it a try! Any and all feedback is welcome.
