Event Analysis Dashboard

A comprehensive dashboard for analyzing presentation interactions and events with filtering capabilities and time-series visualizations.

Features

📊 Event Analysis

• Event Count Over Time: Line chart showing the number of events occurring over time
• Distinct Properties Count Over Time: Line chart showing the count of unique properties (event types, slide types, audiences, slides) over time
• Interactive Time Granularity: Choose between hourly, daily, or custom time intervals

🔍 Advanced Filtering

• Event Type: Filter by interaction source (opinion, answer, reaction, qa_submission)
• Slide Type: Filter by slide type (Poll, Brainstorm, Open Ended, Pick Answer)
• Presentation: Filter by specific presentation names
• Team: Filter by team names
• Audience: Filter by audience names
• Reaction Type: Filter by specific reaction types (like, heart, wow)

📈 Key Metrics

• Total Events count with delta indicators
• Unique Audiences count
• Unique Slides count
• Event Types count
• Peak events and properties analysis
• Average events and properties per time period

📋 Data Preview

• Filtered data table showing relevant columns
• Real-time updates based on applied filters
• Export capabilities through Streamlit

Installation

1. Install Python dependencies:

   pip install -r requirements.txt

2. Run the dashboard:

   streamlit run event_analysis_dashboard.py

3. Open your browser and navigate to the URL shown in the terminal (usually http://localhost:8501)

Usage Guide

1. Data Selection

• Use the sidebar to select from available data files:
  • Sample Data (sample_presentation_data.csv)
  • Dave Presentation (dave_presentation_interactions.csv)
  • Duke Presentation (duke_presentation_interactions.csv)

2. Applying Filters

• Event Type: Choose specific interaction types (e.g., only show poll responses)
• Slide Type: Filter by slide categories (e.g., only show quiz slides)
• Presentation: Focus on specific presentations
• Team: Analyze team-specific interactions
• Audience: Focus on specific audience segments
• Reaction Type: Analyze specific reaction patterns

3. Time Analysis

• Time Granularity: Select how to group time intervals:
  • 1H: Hourly analysis
  • 2H: 2-hour intervals
  • 4H: 4-hour intervals
  • 6H: 6-hour intervals
  • 12H: 12-hour intervals
  • 1D: Daily analysis

4. Interpreting Charts

• Top Chart: Shows event frequency over time
• Bottom Chart: Shows diversity of properties over time
• Peaks: Indicate high-engagement periods
• Patterns: Help identify trends and user behavior

Data Structure

The dashboard works with CSV files containing the following key columns:

• Createdat: Event timestamp (primary time column)
• Interaction Source: Type of interaction (opinion, answer, reaction, qa_submission)
• Slidetypenormalized: Normalized slide type
• Slidetitle: Title of the slide
• Audienceid: Unique audience identifier
• Audience Name: Name of the audience member
• Team Name: Team identifier
• Presentation Name: Name of the presentation
• Reactiontype: Type of reaction (if applicable)

Customization

Adding New Filters

To add new filterable properties, modify the get_event_properties() function in event_analysis_dashboard.py:

def get_event_properties(df):
    properties = {}
    # Add your new property here
    if 'Your_Column' in df.columns:
        properties['Your Property'] = df['Your_Column'].unique().tolist()
    return properties

Modifying Charts

To customize the visualizations, edit the create_charts() function:

def create_charts(event_counts, distinct_properties, filters_applied):
    # Modify chart appearance, colors, or add new chart types
    pass

Troubleshooting

Common Issues

1. Data Loading Errors

   • Ensure CSV files are in the same directory as the script
   • Check file permissions and format
   • Verify date column formats match expected patterns

2. Chart Not Displaying

   • Check if timestamp columns contain valid dates
   • Ensure filtered data is not empty
   • Verify time granularity selection

3. Performance Issues

   • Use smaller time granularities for large datasets
   • Apply filters to reduce data size
   • Consider sampling data for very large files

Data Format Requirements

• Timestamps: Must be in format DD-MM-YYYY, HH:MM
• CSV Encoding: UTF-8 recommended
• Missing Values: Handled automatically with 'unknown' or 'N/A' placeholders

Advanced Features

Export Functionality

• Use Streamlit's built-in data export features
• Right-click on charts to save as images
• Copy filtered data to clipboard

Real-time Updates

• Dashboard automatically refreshes when filters change
• Cached data loading for better performance
• Responsive design for different screen sizes

Contributing

To contribute to this dashboard:

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Test with sample data
5. Submit a pull request

License

This project is open source and available under the MIT License.