ML Interpretation Dashboard

Interactive visualization of feature importance using SHAP

As machine learning models become more complex, interpretability becomes critical for trust and deployment. I built and deployed an interactive dashboard designed specifically to demystify model predictions using SHAP (SHapley Additive exPlanations) values.

The project features a full-stack implementation:

  • Backend: Developed end-to-end APIs utilizing FastAPI to handle model training, evaluation, and data serving.
  • Frontend: An interactive interface allowing users to explore feature importance visuals dynamically.
  • Deployment: The entire system is fully containerized using Docker, ensuring consistent environments and scalable deployment.
<!--
  See https://www.debugbear.com/blog/responsive-images#w-descriptors-and-the-sizes-attribute and
  https://developer.mozilla.org/en-US/docs/Learn/HTML/Multimedia_and_embedding/Responsive_images for info on defining 'sizes' for responsive images
-->

  
    <source
      class="responsive-img-srcset"
      
        srcset="/assets/img/dashboard_ui-480.webp 480w,/assets/img/dashboard_ui-800.webp 800w,/assets/img/dashboard_ui-1400.webp 1400w,"
        type="image/webp"
      
      
        sizes="95vw"
      
    >
  

<img
  src="/assets/img/dashboard_ui.jpg"
  
    class="img-fluid rounded z-depth-1"
  
  
    width="100%"
  
  
    height="auto"
  
  
  
  
    title="Dashboard Interface"
  
  
  
    loading="lazy"
  
  onerror="this.onerror=null; $('.responsive-img-srcset').remove();"
>

</picture>

</figure>

</div>

</div>

The primary dashboard interface, rendering real-time SHAP dependency plots and summary visualizations.

–>