Dashboard Layout and Data Source Improvements

[TeachMeTW]

So the ml dashboard looks like this right now:

Current Behavior

• Dashboard appears blank until a CSV file is loaded
• Tools and parameters are in the sidebar, reducing discoverability
• Requires manual CSV upload/specification for data input
• Limited user interaction when first opening the dashboard

Proposed Changes

1. Layout Restructuring

• Move key tools and parameters from sidebar to main interface
  • Create a dedicated "Settings" tab or collapsible section in the main area
  • Organize parameters into logical groups with clear labels
  • Add tooltips/help text for better parameter understanding

2. Initial State Enhancement

• Display welcome message/getting started guide on first load
• Show sample visualizations or dummy data until real data is loaded
• Add quick-start tutorial or walkthrough for new users
• Include parameter presets for common use cases

3. MongoDB Integration

• Replace CSV dependency with MongoDB connection
  • Add MongoDB connection configuration section
  • Implement data collection dropdown/selector
  • Add date range filters for data querying
  • Include real-time data refresh capability

4. UX Improvements

• Add loading states and progress indicators
• Implement error handling for database connection issues
• Add data preview before processing
• Include "Save Configuration" feature for frequent use cases

Technical Requirements

• MongoDB driver integration
• Data streaming/pagination for large datasets

Impact

• Improved user experience with clearer interface
• Reduced friction in data loading process
• More robust data source management
• Better scalability for larger datasets

Acceptance Criteria

• Dashboard shows meaningful content on initial load #37
• All parameters accessible from main interface
• Successfully connects to MongoDB
• Loads and displays data without manual file upload
• Maintains current functionality while improving UX
• Includes proper error handling

Notes

• Remember our user base -- make it dummy friendly