What it does
Meet Wiki-pipeda bot - a nearly autonomous bot built on the Keybase chat API that helps you determine what's the best online service that will respect your privacy. The in the works web crawler listens for changes in privacy terms within major providers, and aggregates their care about privacy with respect to PIPEDA with a single source of truth: a number from 0 to 10. The closer a score to 10, the better a service will likely respect your privacy.
How we built it
We worked with the KeyBase API and built out a small server from Python. With this we created our own complex but logically correct algorithm for calculating what factors in a company's wording create a warning sign for privacy and which stand in defense of it.
Challenges we ran into
There were three major challenges that we overcame throughout this project. The first was working on actually improving the KeyBase API - we worked to create a new method that allows file transfers to the BOT! The second was our initial obtaining of data from the internet - unstructured data that has to be parsed in an equal way is quite difficult and we had to elect to use some strict heuristics. Lastly was our choice of how to appropriately represent PIPEDA - focus on which pillars, what not? We believe our work sets beginning research into future modules building upon this.
Accomplishments that we're proud of
Our first ever chat bot implementation
Learning of PIPEDA as well as creating a potential standard for representing it online
What we learned
Working with chat bots is fun and something we'd love to continue to work with the KeyBase API. Working with web crawling techniques in an on-premise scenario is hugely rewarding.
What's next for Wikipipeda
Version 0.1: developing a cloud native app that retains the same familiar interaction with Keybase users, and is able to scrape more data for better measurements.
Built With python domain: wikipipeda.tech