Inspiration
Inspired by the inefficiency of long convoluted instructions, we sought to create something that allows maximum productivity by transforming the dauntingly complex into simple, user-friendly guides.
What it does
HowTo is able to dissect a story, tag each word by assigning it its respective part of speech, and output a diagram of instructions with arrows indicating the flow.
How we built it
We used the dependency parser to extract necessary pieces from instructions. Then we displayed each instruction in a block of it's own using graphviz with arrows showing the flow.
Challenges we ran into
Extracting relevant parts of the text, removing irrelevant parts, and working with inconsistent python packages and distributions.
Accomplishments that we're proud of
We are proud of understanding how data representation helps solve problems, extracting relevant information from the complex parse tree generated by Stanford parser, working together despite all frustrations. Additionally, we are proud to have created a viable product with practical application.
What we learned
How NLP could be imperfect and unintuitive Python packages can be.
What's next for HowTo
We will work to incorporate a way to include secondary (extra) information that makes the main instruction clearer as fluidly as possible. Furthermore, HowTo can be extended to fields such as education, medicine, business, and intelligent services.
Log in or sign up for Devpost to join the conversation.