Amazon Dash Recycle

LA Sanitation partners with Amazon to develop recycling-related processes and products.

Comment 5

Inspiration

With the directive to help reduce contaminants in the recycling process and to use the provided data, we determined that the best path towards success would be to partner with Amazon to help reduce some of the common contaminants through education. Ours is a three-pronged approach:

  1. Education -- Leverage existing Amazon Give Back Box program -- City of LA would provide Amazon with recycling postcards and recycling tips that relate specifically to products they sell: Food and Clothing in particular -- Urge Amazon to include recycling recommendations post-sales for items that commonly contaminate: eWaste, clothing, and food contamination -- Develop an Amazon Alexa Skill for recycling answers.
  2. Low-Tech Entry Point -- Partner with Amazon to develop an Amazon-DASH-style device that goes inside recycling bins and tracks opens/closes of bin, and takes pictures that are sent to Amazon Cloud
  3. High-Tech Entry Point -- Use images with machine learning to process contents of bins and determine rates of contamination, theft/scavenging, recycling

What it does

  • Amazon Notification Services add recycling recommendations to order confirmation, out-for-delivery, and delivery notices. "Your Amazon order of TP-Link Router has shipped! Click here for recycling info for this item or it's packaging." "Your Amazon order of Nike Running Shoes has been delivered. Click here for how to recycle our packaging."
  • Amazon Alexa skill answers questions about where items go, how to properly recycle them, and where they can get more info. The information can be based on ZIP code inclusion, and jurisdictions can submit their XML instructions for inclusion in the larger data set. "Alexa, is clothing recyclable?", "Alexa, how do I recycle cardboard?"
  • The Amazon Dash Recycle - Small, camera-enabled device that piggybacks off of the Amazon DASH technology to send open/close data and images to Amazon Cloud Services.
  • Amazon Cloud Services - Use image processing to determine when packaging goes to recycling. Determine when "bad" items are thrown away... "Amazon Dash Recycle noticed that you put a garden hose in your bin. Would you like to know how to properly dispose of this item?"

How we built it

  • Amazon Notification Services API - would add links to recycling information for items that can/should be recycled
  • Amazon Alexa Skills API - creates framework for questions that can be asked
  • Amazon Cloud Services - leverage image processing APIs to determine contents of images

Challenges we ran into

  • Developing the Recycle Dash to make it affordable.
  • Getting Amazon to partner

Accomplishments that we're proud of

  • Amazon could provide the City with multiple outreach opportunities at low cost.

What we learned

  • Textiles and Food/Grease are the two biggest contaminants in LA's trash stream - making up about 60% the total contamination.
  • Amazon has 43% of US online retail sales. Clothing sales are one of Amazon's fastest growing categories, and they are looking to dominate the food industry with the purchase of Whole Foods and the rollout of Prime Pantry.

What's next for Amazon Dash Recycle

  • Continue to monitor contamination and dial in messaging using google's data.
  • Amazon's AI will continue to improve as more recycling is photographed.

Built With

Share this project:

Updates

posted an update —

Image Tagging: Automatically Searching for Contaminants

Once pictures of the bin contents are taken the images will be analyzed by Microsoft's computer vision API. It takes the images as input and returns "tags" as output, where tags are objects found in the image.

Once an image is tagged the tags can be scanned against the list of contaminants. 1.Outreach to residents/regions with high levels of contaminants. + Send resident an email. + Attach "oops tag" to bin. + Deduct from recycling incentive

Log in or sign up for Devpost to join the conversation.