Inspiration
From all the provided HackDavis prompts, Greenhouse's struck out the most. Because most of us have experience with machine learning and AI, the prompt laid well in our area of expertise and we believe we could tackle Greenhouse's problem via neural networks and classification.
What it does
Our algorithm takes an input book pdf or document pdf, analyzes it, and outputs the appropriate reading level of that pdf.
How we built it
We coded a simple neural network and classifier that analyzes pdf readability via word count and word length.
Challenges we ran into
A lot to be honest. Python has a readability library available, but for some reason, it wasn't working out when we set it up. We ended up using a library called PyPdf2, though the problem with that is that it could only analyze simple documents, so long, complex books were giving problems. We only have 2 variables, so it won’t be as effective.
Accomplishments that we're proud of
We were able to make an algorithm that analyzes a pdf's reading level.
What we learned
One big thing we learned is the complexity of a neural network. It's like a rabbit hole, the weights and variables can get complex, and it's hard to fine-tune it.
What's next for Reading Level Analyzer
Just living life man. Hopefully graduate in June too.
Log in or sign up for Devpost to join the conversation.