We were interested in how many tech buzz-words are used in tech start-ups so we wanted to see if we could generate a website for a fake one that was indistinguishable from a real one.

What it does

It generates a company name and headings using markov chains

How we built it

The markov chains are trained using text we scraped from about 50 tech startup websites. Then we used a java servlet to import the text onto a website we made. We used a python script using beautiful-soup to compile the headings from 50 tech startup websites

Accomplishments that we're proud of

It works pretty smoothly considering the short amount of time


Gathering lots of useful data was difficult. If we just scraped all the text from the website we would get lots of useless data that wasn't "buzzwordy" enough. So we ended up going with all the headings from the websites. Since we were new to web scraping, learning this process was difficult

What's next for CyberDuck2077

Generating a whole entire company

Built With

Share this project: