RescueLab Hackathon 2023

Team 1. Scraping content from a web app. The goal here is to build a tool that grabs content from an app and saves it to a database (or a local file). Many apps are also available through the web, which makes scraping easier. In fact, for the most popular apps, there are packages that allow scraping (e.g., Instagram). There are a variety of software packages that can help in general. 

Moderate Difficulty. This task will require programming experience. You'll work with a team and focus on a specific web-available app. Your goal will be to scrape a lot of sample data via code and store it in a format that is compatible with the labeling Team 3's work.

Team 2. Scraping content from a mobile app. It's harder to scrape when there is not a website for a mobile app. (For example, you can use Instagram via an app or the web.) In these cases, it's possible to scrape by intercepting traffic as it is received by the app using MITM Proxy. A tool like Appium can automatically scroll the screen to ensure content keeps arriving to scrape.

More than Moderate Difficulty. This task will require more programming experience, some knowledge of networking, and the willingness to monkey with things. As mentioned on the left, you'd probably use MITM Proxy and Appium. As above, your goal is to pick an app and scrape some content by the end of the day suitable for Team 3 to use. 

Team 3. Labeling content. We need humans to examine a variety of content and label it according to the age ratings that Google and Apple have provided. You can let us know ahead of time if you want to label content that is Safe for Work, or if you are ok with labeling content that one would call Not Safe for Work. 

Not Difficult.  This task will not involve programming. We'll spend some time reading the age rating guides, agree on a labeling system, and then start labeling. We'll use Label Studio to do the work. A big part of the job will be keeping everything organized. Your goal is to give data to Team 4 and 5 to use. We'll have some content ready to label when you walk in, you don't have to wait for Teams 1 or 2 to get started. 

Team 4. Training a Text or Image Classifier. With labeled examples in hand, we need folks to train a classifier to identify the age level of content. We plan to handle text and images distinctly. 

Expert Difficulty. To take on this task, you'll need experience with training a classifier. We'll give you labeled data to start the day so that you don't need to wait for Team 3. One group will work on classifying text, another on images. If you are brave, you could work on a multimodal classifier. 

Team 5. Wrangling Apps for Scraping. Some apps put up hurdles to scraping. For example, they might be "pinned" to a specific public key and they need a nudge in the right direction. 

Expert difficulty. To take on this task, you'll need experience unpinning apps and with reverse engineering. If you are successful, you'll be able to hand your work off to Team 2.