Inspiration
30% of items in recycling bins aren't actually recyclable. This means that we have to ship our recycling to other countries, increasing costs and greenhouse emissions. This should change.
What I did
I started the process of analyzing data from recycling bins to understand what, when & where we should focus outreach efforts to reduce contamination in recycling bins.
What I found
Look at presentation that is linked. Exploratory graphs + description in notes make the relevant points
What's Next for Containing Contamination
Scale up
- Group recycling materials by cost of recycling: Some items cost more to recycle than others. Grouping would allow us to predict how cost of recycling would change with time. We could also have seasonal models by type of material
- Clean data from before 2013 : Current data doesn't have enough data points to make strong statements. 2005 - 2013 data exists. Should take a couple hours to clean and include.
Past Outreach Study - CLEAN CLEAN CLEAN. We can target our outreach efforts better by answering the following questions:
- What is the distribution of contamination?
- Does the type of contaminant vary by region? A quick look reveals that in the
New Data Collection
- Redo past outreach with better data collection practices. (Examples (1) csv files with tidy data principles followed (2) Lots of stuff to say here. Not enough time to type it
- Map areas & subareas to zipcodes. Because US zipcodes mirror socioeconomics, we can use this mapping and use small pilot studies to make predictive models for type of contamination by area.
- Which contaminants have the worst cross-contamination effects? Contaminants mess up other recyclable stuff to. Some contaminants do this worse than others. We should focus outreach efforts on these contaminants and areas that produce
- Collect data on trash (not recycling) cans:- how many recyclable materials being thrown in regular trash? So far we have looked at the contents of recycling bins to understand contamination. We don't have a sense of how far away we are from maximizing recycling output and where?
What I learned
Don't spend 4 of 6 hours at a hackathon trying to clean a dataset
Log in or sign up for Devpost to join the conversation.