-
-
Home page
-
Results page
Inspiration
Fast fashion is becoming a well-known phenomenon that impacts the environment negatively. Approximately 11 million tons of clothing is thrown out in the United States. Fast fashion does not only relate to environmental impact but human rights as many fast fashion retailers employ employees in third world countries with low wages and terrible working conditions. Our aim with Sustainable Style is to try to minimize the waste going into the environment as well as subtlety promoting good working conditions and higher wages for workers.
What it does
Users upload a photo of a clothing article they like that they find online, whether from a fast fashion retailer or even a sustainable brand, and Sustainable Style will then generate a list of various alternative and sustainable options from sustainable brands for the user that is similar to the photo they uploaded. Other than sustainable brands, Sustainable Style will also use secondhand thrift stores. Secondhand clothing, no matter the brand, helps the environment because instead of the clothing article going to waste, another person will be able to own and wear it, creating a recycle reduce, reuse, recycle effect.
How I built it
This web application relies on the Google Cloud Vision Product Search API to match fast fashion items with similar sustainable products. As of right now, there is a small data set with various sustainable items stored in a CSV file in Google Cloud Storage. In order to avoid filling the suggested items with random items, the items displayed are sorted from highest confidence to lowest confidence. The confidence scores are based on how certain the API is that the items are a match. Generally, a match of 35% or above is an indication of a likely match.
API's used:
- Google Cloud Vision Product Search
- Google Cloud Storage
Challenges I ran into
Although there were examples of how to use the Google Cloud Vision API to perform basic operations, documentation beyond the basics were scarce. The Google APIs used many classes that had different ways of accessing member values which were necessary. When querying items, not all the fields that were provided were returned, and this was also not very well documented.
Accomplishments that I'm proud of
We successfully came up with an idea to tackle one of the verticals at HackSC 2020 and built a complete application by using Flask and various Google Cloud API's.
What I learned
In 36 hours, we learned how to use one of Google Cloud's machine learning API's in our project. We were able to provide the API with the data that is necessary for the API to correctly determine similarities between images of fast fashion and sustainable products. We also learned how to build a web app from scratch, from creating the front end and back end individually to connecting the two code bases with Jinja and Python.
What's next for Sustainable Style
In the future, we would like to create a mobile app so users can shop from their cell phones. In addition, we would like to have a sustainability scale, in which users can choose how sustainable they want their clothing. The sustainability scale will essentially factor in the environmental impact a brand has as well as if their employees are in good working conditions. We would also like to implement scales such as Price (Low to High) or Price (High to Low). Also, the data set would be improved so that there are more sustainable items that can be matched to. In the future, this app could be deployed to Google App Engine, and all its functions would be able to be done exclusively through Google Cloud Storage.
Log in or sign up for Devpost to join the conversation.