Recycling cans is a huge source of income for many people that displaced or homeless; they might make up to $70 in a single day. However, looking for cans and can-dense areas can be challenging, especially for some that are more physically disabled, thus making them less efficient. In addition, upon receiving their payment, many homeless people will spend their money unwisely, not considering a long term plan.
What it does
To assist our friends, we built sCANvenger that does two things:
1) It generates predictive models of aluminum can concentrations, depicted as a heatmap. We can guide users to areas that have the most cans, and with the least amount of pickers. 2) The ecosystem also automatically deposits into a Capital One bank account, created for pickers using our system for the first time. Depositing and withdrawing money comes with advice to encourage smart spending.
We envision our application being primarily used in two locations, respective to the functionalities described above:
1) By placing a large screen monitor in a homeless shelter, we hope that having a source of information that optimizes their day and income flow will also encourage seeking out shelter as opposed to living on the streets.
2) We hope to create a partnership with recycling centers, which would have a way to access our recycling center/picker portals, which allow for depositing their income into secure bank accounts as well as providing information about their money flow.
How I built it
As it is hard to build such a system without adding a lot infrastructure (e.g. smart trash cans) to keep track of number of aluminum cans in different areas, we take advantage of other accessible data sources such as population density, supermarket distributions, and household distributions. With the power of HTML5's geo-location, we are able to locate the device's current geological location. By mining data from heterogeneous data sources and using statistical inferences based on demographic models, we are able to infer the density of aluminum cans over an specific area as well as the amount of money a picker can make in that area per hour.
We then used Capital One's Nessie API to facilitate bank transactions.