StrokeSense

Built a React dashboard that visualizes stroke risk using interactive heatmaps, boxplots, scatter plots, and line charts. Our model predicts stroke risk based on age and symptoms for better insights.

Comment

Inspiration

Stroke is one of the leading causes of death and disability worldwide, and early detection of stroke risk can save lives. We wanted to create a data-driven approach to help visualize and predict stroke risk based on symptoms and age.

What it does

StrokeSense is a data dashboard that visualizes stroke risk factors using interactive visualizations. It includes:

  • A heatmap that shows the correlation between symptoms and stroke risk.
  • Boxplots comparing the risk of having symptoms vs. not having them.
  • A scatter plot displaying age vs. stroke risk distribution.
  • A line plot showing symptom distribution by risk category.
  • A predictive model that takes age and symptoms as input and predicts stroke risk as either high or low.

How we built it

We developed the frontend UI using React to create an intuitive and interactive dashboard. The backend was built with Python using FastAPI to handle data processing, model training, and API responses. The dashboard visualizations were implemented using Plotly to dynamically display data insights. Our predictive algorithm leverages a RandomForestClassifier trained on symptom and age data to classify stroke risk levels.

Challenges we ran into

  • Finding a reliable dataset that covered a wide range of stroke-related symptoms.
  • Optimizing model performance while maintaining fast response times for predictions.
  • Designing a dashboard layout that effectively communicates insights without overwhelming the user.
  • Ensuring the FastAPI backend integrated smoothly with the React frontend.

Accomplishments that we're proud of

  • Successfully integrating an interactive dashboard with meaningful visualizations.
  • Developing a predictive model that provides real-time stroke risk assessments.
  • Optimizing data processing to handle large datasets efficiently.
  • Creating a user-friendly experience that makes complex medical data more accessible.

What we learned

  • Improved our skills in React for frontend development and FastAPI for building efficient backend APIs.
  • Learned how to optimize machine learning models for better real-world application performance.
  • Gained insights into effective data visualization techniques for medical data.
  • Understood the importance of designing user-friendly interfaces for health-related applications.

What's next for StrokeSense

  • Improving the predictive model by incorporating more features such as lifestyle factors and family history.
  • Enhancing the dashboard with additional interactive visualizations.
  • Deploying the project online to make it accessible to a wider audience.
  • Partnering with healthcare professionals to refine the accuracy and usability of our tool.

Built With

Share this project:

Updates