Text2Test

Inspiration

Professor Michelle McAllister, formerly at Northeastern University, provided notes styled in such a manner. These "incomplete" notes were filled out by students during her lectures. A team member was her student and, having done very well in her class, wanted to create similarly-styled notes for other classes. The team member envisioned other students being able to style their notes in such a way so they could practice for tests and ace their exams!

What it does

By entering their notes to a textbox, students can use Text2Test to create fill-in-the-blank questions in order to practice for tests.

How we built it

We used AWS APIs, specifically Lambda, Comprehend, Gateway. Additionally, we used HTML, JavaScript, Python and Ruby.

Challenges we ran into

We had more ideas we wanted to implement, like running Optical Character Recognition (OCR) on an uploaded PDF scan of handwritten notes, but we would need more than 24 hours for such an ambitious idea. Having never used AWS prior, we had to learn how to upload code to the Lambda API and connect the Comprehend and Gateway APIs to it. We also had difficulty posting requests from the website to the Gateway API, but we eventually solved that problem. The Python code for Lambda code and the website code are saved to GitHub.

Accomplishments that we're proud of

This website allows students to create a profile and login to a portal. Students can schedule exams, upload their notes via a submission portal, add notes to a particular subject, generate customized tests based on their notes, and test their knowledge by entering answers to fill-in-the-blank questions.

What we learned

We learned how to utilize many AWS APIs like Comprehend, Lambda, and Gateway. In addition, we learned how to post requests in the right format from outside entities (like our Ruby on Rails website) to the AWS Gateway API.

What's next for Text2Test

We want to give students the ability to upload handwritten notes. To increase the numbers of learning tools, we wish to include a multiple-choice section with "smart" answer choices. Finally, we want to allow images to be used as inputs and test students on content pertaining to the images.