Inspiration
The name SmartUP With Tech means just that: we're leveraging technology to smart concepts up for students, rather than dumbing them down. Since the pandemic, students (particularly those without equitable access to resources) have been relying on online resources for critical components of their education. Without a guiding hand, these students have had significant barrier to learn difficult subjects, discouraging them from learning. Students have also commonly faced this problem where the information they receive in their classroom is well above their comprehension level. We wanted to empower these students and any learners looking to gradually increase their knowledge.
What it does
SmartUP! Takes a search query, as you'd type in Google, and returns you something more intelligent: web results that have been intelligently scanned and sorted by the reading difficulty of the material.
How we built it
We constructed a react front-end that communicates with a nodeJS backend and fetches relevant data for our webquery API call, website scraper, and article evaluation algorithm. The algorithm in question is one of the most widely-used metrics for reading difficulty, the Flesch-Kincaid algorithm.
For user authentication and user-specific data, we used global firebase which will communicate to our React app to improve the quality of recommendations provided to our students. Finally, we used domain.com to register our "smartupwith.tech" domain which we intend to host using heroku.
Challenges we ran into
Half of our team is new to hackathons so we had a lot of difficulty getting everyone onboard full-stack web-development. Getting the team on the same level was a hurdle that we encountered and out-smarted.
Additionally, we envisioned a wider scope for our project but as we progressed, we made concessions as we crystallized our vision. Namely, -We chose a simpler algorithm for ranking our articles -Used less data for the algorithm -A less streamlined website
Technical challenges were abundant in this project. Mohamed had difficulty figuring out the best way to make API calls from a dueal React-Node application, and CORS errors haunted him for the night. Eli delved into React and Node for the first time ever, and this involved understanding how requests are processed in the back end. Param dealt with synthesizing the various (out of sync) branches together and allowing data to be passed through many layers of DOM. Sid worked on the recommendation engine and figured out the most feasible algorithms to use.
Accomplishments that we're proud of
We're proud of the idea we have and what it means. Education is idealistic because in its face, all are equal, all are humbled. But in reality, education fails to be an equalizer, if anything it only exacerbates differences in social conditions and resources.
For many people, education is their opportunity to escape their current conditions and is commonly cited as the primary vehicle of upward social mobility. By providing a tool that can improve the educative experience, especially for those of depressed socio-economic conditions, we help the world globally.
We are also proud of the functional output of our website, considering how long we worked overnight with something new constantly breaking. The GitHub logs of commits to this project show dedication, perseverance, and love for creating. The final product is looking promising for the future, as well: our pride stems from the polish and raw usability that our website presents users. Users are faced with a simple, user-friendly interface without distracting bells and whistles.
What we learned
We are so proud of the what we learned in the process of bringing our idea to fruition. The novice members of our team learned much about the agile development methodology, git, and React. As a team, we learned how to integrate Google's account authentication services, firebase, and most importantly how to work together.
What's next for SmartUP!
Moving forward, we look to advance our backend algorithms and collect more informative user data, featuring more advanced machine learning, like using a partial bandit feedback recommendation algorithm; create a custom database of user-curated articles; and establish an active user base. As college students, we also hope to someday develop a PDF aggregator that can perform our algorithm on research articles, that often take a very technical route.
Log in or sign up for Devpost to join the conversation.