Project Descriptions

🌐 User Site

Objective: Our user site allows Berkeley students to request associate compiled resources and view other C@C resources such as Bear Buddies.
Key questions:
- How can we display and convey all of Connect@Cal's services to students?
- How can we solidify Connect@Cal's branding through the main site?
- How can we ensure retention and decrease bounce rate of old and new users?
We are currently undergoing a redesign/rebranding of the site and aim to finish building the site by this semester.
For this project, we are working with HTML/CSS, Javascript, and Bootstrap.
Progress so far:
- Figma Mockup
https://www.figma.com/embed?embed_host=notion&url=https%3A%2F%2Fwww.figma.com%2Ffile%2FbA5lm8MFKXEv1Q6cIFWOdd%2FC%2540C-User-Site-HiFi-mockup%3Fnode-id%3D0%253A1
- (Work in progress) https://github.com/Connect-At-Berkeley/redesigned-website
- Fall 2021 Tasks:
  - Partners and Impact, Donate, FAQ, Home, Our Story pages
  - Animation + interactivity

📚 Resource Finder

Objective: Using the Resource Finder, users will be able to easily access resources surrounding the UC Berkeley community.
For this project we address these key questions:
- How can we better automate resource providing for students?
- How can we make resource finding as seamless, efficient, and helpful as possible?
- How can we make our resources open source but also user-friendly?
Currently, users can filter using "Identity" and "Issue Area" dropdowns or do a keyword search, and if it’s in our database, the website will display a list of resources in the format of a gallery view.
For this project, we are working with React (frontend) and Flask (backend) to build our web application, and deploying it on OCF and AWS. Knowledge about these tools and technologies is definitely a plus but not required!
Progress so far:
- Figma Mockup:
- (Work in Progress) Resource Finder: https://connect-resource-finder.herokuapp.com/
Fall 2021 (Tentative) Tasks:
- Making UI Changes
- Syncing, building, and scaling our Resource Database
- Adding Image Scraping
- Adding Crowdsourcing Functionality
  - A student can submit a resource to our database, which automatically shows up in the Resource Finder
- And more!

🧭 Navbuddy

Objective: Using NavBuddy, Operations Associates will be able to easily update students with new resource matches found for their past cases.
For this project we address these key questions:
- How can we update students with a continuous stream of useful resources?
- How can we allow Ops Associates to easily identify relevant cases for every new resource?
- How can we improve our resource provision with user and Ops feedback?
Currently, Ops Associates can view a "queue" of relevant resource-case matches which they can use to draft emails with new resources to students who previously submitted cases. This queue of new matches is created using a custom bag of words and cosine similarity algorithm which matches new resources with relevant past cases.
For this project, we are working with Flask and jQuery to build our web application, and deploying it on Heroku. The algorithm used to match cases and resources is implemented with scikit-learn. Knowledge about these tools and technologies is definitely a plus but not required
Progress so far:
- Simple Heroku webapp with naive UI:
- Naive matching algorithm which finds any cases to which a new resource may be relevant, using a simple bag of words and cosine similarity algorithm.
Fall 2021 (Tentative) Tasks:
- Investigate better resource-case matching algorithm
- Migrate to AWS
  - Matching algorithm as AWS Lambda job
  - Host webapp on EC2 with its own domain
- Allow Ops to mark matches as good or bad
- Improve table and frontend UI
- Add email templates to allow Ops to directly draft emails via NavBuddy
- And more!