Ideally, a web scraper for tracking elements robust to website changes. Currently backend parses the DOM and uses simple heuristics on class/id/hierarchy to pick the most likely target. chrome-ext is a chrome extension that allows you to pick web page elements to create the website's api.

More interesting directions/cues:

• What kind of value are trying to track
• What is the typical layout of websites created with framework X?
• How is content presented/highlighted through common design guidelines
• What does the web platform actually do? level of interaction?