Inspiration
High school students who are enrolled in multiple AP courses tend to face heavy academic workloads with several assignments, projects, and exams. As workload increases, many students struggle with effective time management, leading to procrastination, delayed bedtimes, and therefore, chronic sleep deprivation. This issue is also prevalent in my high school as well, and it inspired me to develop an AI-based solution that might address this lack of sleep among high schoolers.
What it does
The SleepWise platform analyzes assignment data, deadlines, and each student's historical work pattern to estimate how long future assignments will take for that student. Using these personalized predictions, SleepWise also identifies a sleep risk percentage caused by the workload and generates optimized study plans that help students complete their work while maintaining healthy sleep habits. In addition, the platform takes feedback from students of actual completion times so the AI can effectively adapt to their work pattern over time.
How we built it
Sleepwise was built as an AI-powered prototype using Microsoft Copilot, where we designed a multi-layer architecture that incorporates data input from Google Classroom API, which provides assignment details, deadlines, and course information. As real student data was not used, we generated synthetic datasets that simulate AP student workloads, study habits, and assignment completion patterns. The system would then process this data from a backend layer built with container apps and functions that standardize and prepare the information for analysis. The main AI processing of SleepWise is powered by Python-based ML models, which includes a predictive regression model that estimates assignment completion times based on historical work patterns for each student. These predictions are fed into a recommendation system and optimization engine that generates personalized study schedules and calculates sleep risk (0-100%). In the end, the outputs are delivered in the user-interface dashboard that presents students with workload predictions, optimized study plans, and proper sleep recommendations.
Challenges we ran into
One potential challenge in developing SleepWise is that the accuracy of the AI-generated workload predictions relies heavily on the quality & consistency of historical student data. If students do not regularly input accurate completion times or if their work patterns tend to vary significantly due to other outside factors, the model may produce unreliable predictions, which can reduce the effectiveness of the recommended study plans.
Accomplishments that we're proud of
A major accomplishment that our team is especially proud of is how through the development of SleepWise, we've created an opportunity for students to effectively manage their academic workload with minimal pressure and stress. Most importantly, students would now have access to AI-suggested work plans and can receive quality sleep while ensuring optimal studying.
What we learned
Through this project, we learned the core AI capabilities and the responsible AI guardrails associated with the creation of SleepWise. Specifically, we came to know about the predictive workload regression model, and how this model is essential for predicting workload patterns and thereby generating optimal study plans. Additionally, we realized that there exists specific AI guardrails for this project, like fairness, reliability, accountability, transparency, etc., and these aspects are crucial for minimizing risk in the system.
What's next for SleepWise
Our next step would be to conduct a pilot study with approximately 10-20 AP students, such as students from our own high school, to evaluate how accurately SleepWise predicts workload and supports time management in the real-world. We can then use the collected feedback and completion-time data to further refine the predictive models and improve recommendation quality. After validating the system, we would expand integration with additional educational platforms and explore partnerships with schools to reach more students.
Log in or sign up for Devpost to join the conversation.