7 videos • 📅 2024-11-21 09:00:00 Africa/Ceuta
Watch Video
1:38:36
2024-11-21 09:42:06
18:30
2024-11-21 11:49:40
2:57:14
2024-11-21 13:02:19
2:10:52
2024-11-22 09:10:59
1:51:09
2024-11-22 09:32:15
56:00
2024-11-22 13:05:24
1:24:05
2024-11-22 14:24:08

Course recordings on DaDesktop for Training platform

Visit NobleProg websites for related course

Summary

Overview

This course session provides an in-depth, hands-on exploration of Apache Superset for data visualization and dashboarding, with a focus on dynamic query personalization, public access configuration, geographic visualizations (GeoJSON, deck.gl, country maps), and integration with external data sources like Google Sheets. The trainer demonstrates advanced features including user-based macros, URL parameter injection, feature flags for public role access, custom map integration, and interactive dashboard layout design. The session includes live troubleshooting, collaborative problem-solving, and practical exercises on geospatial data handling, culminating in a guided workflow for creating and sharing interactive dashboards with both authenticated and anonymous users.

Topic (Timeline)

1. Dynamic Query Personalization & User Context [00:00:01 - 00:13:47]

  • Demonstrated use of SQL queries that dynamically inject user context using Superset macros: {{ current_user_name }}, {{ username }}, {{ transaction_id }}.
  • Showed how to retrieve browser/environment variables (e.g., from URL parameters) directly within queries using Jinja templating syntax ({{ }} for values, {% %} for logic).
  • Highlighted use cases: personalizing dashboards per user (e.g., showing only their data), filtering by URL parameters (e.g., ?department=FR-01).
  • Warned about security risks: unvalidated URL parameters can enable SQL injection; recommended use only in controlled environments.
  • Discussed integration with eFrame (embedded dashboards) to pass parameters securely via URL without requiring login.

2. Public Access & Feature Flags Configuration [00:13:47 - 00:26:40]

  • Explained how to enable public access to dashboards via Superset’s PUBLIC_ROLE_LIKE feature flag, setting it to gamma to grant read-only access to unauthenticated users.
  • Clarified distinction between user roles (Public, Gamma) and feature flags: feature flags are global configuration toggles, not per-dashboard permissions.
  • Demonstrated workflow: (1) enable PUBLIC_ROLE_LIKE=gamma, (2) assign dashboard read permissions to the Public role individually.
  • Showed that public and authenticated users share the same dashboard URL; permalink functionality preserves filter states for sharing.
  • Emphasized that administrators, not end-users, manage these settings.

3. Geographic Visualizations: Country Maps & GeoJSON [00:26:40 - 00:48:17]

  • Introduced three Superset geospatial chart types: Country Map, Deck.gl, and GeoJSON.
  • Focused on Country Map: used built-in France department and region maps; explained required ISO codes (e.g., FR-01) for data joining.
  • Demonstrated how to add custom geographic boundaries (e.g., communes) by modifying countries.ts and adding GeoJSON files in the Superset frontend codebase.
  • Warned about performance issues: large GeoJSON files (>1MB) cause frontend lag; recommended simplifying geometries (e.g., reduce vertex density) for 35,000+ communes.
  • Discussed data preparation: join internal dataset IDs with GeoJSON feature IDs (e.g., via lookup table or data transformation).
  • Mentioned QGIS as a tool for simplifying and exporting GeoJSON for Superset.

4. Advanced Geospatial: Deck.gl & Point/Heatmap Visualizations [00:48:17 - 01:50:17]

  • Demonstrated Deck.gl for rendering point clouds, hexbin grids, and heatmaps from latitude/longitude data.
  • Showed how to visualize point data (e.g., oil wells) using Scatterplot and Heatmap layers, with color and size encoding based on categorical/numerical fields.
  • Explained that Deck.gl renders points independently of background maps; Mapbox tiles can be added via API key (not configured in demo).
  • Walked through importing CSV data via Google Sheets: created sample dataset with latitude, longitude, and count fields.
  • Troubleshot data import issues: inverted lat/long fields, encoding errors, and CSV separator mismatches.
  • Confirmed that Deck.gl supports GeoJSON input via a dedicated column in CSV, but CSV with lat/long columns is simpler for beginners.
  • Highlighted flexibility: 5-minute setup from raw data to interactive map.

5. Dashboard Layout, Custom Elements & Data Integration [01:50:17 - 02:10:45]

  • Demonstrated dashboard layout editing: added Markdown text blocks, titles, dividers, and custom HTML (e.g., icons, images).
  • Explained image storage: custom images must be hosted in Superset’s frontend (requires terminal access); not available via UI.
  • Showed how to use Layout Elements to arrange charts in rows/columns, resize containers, and create custom visual hierarchies.
  • Walked through creating datasets from Google Sheets: shared public link, connected via “Connect Google Sheets,” selected sheet tab.
  • Emphasized best practices: use clean column headers, avoid special characters, ensure correct data types (numeric for metrics).
  • Concluded with a call to action: participants to experiment with their own datasets using the provided Google Sheet template and practice creating country maps and Deck.gl visualizations.

6. Wrap-up, Q&A & Next Steps [02:10:45 - 02:10:45]

  • Summarized key takeaways: dynamic queries, public access via feature flags, geospatial data handling, and layout customization.
  • Encouraged participants to test workflows independently using the shared Google Sheet and Superset instance.
  • Announced next session at 13:00, following a break.

Appendix

Key Principles

  • Dynamic Queries: Use Jinja macros ({{ current_user_name }}, {{ url_param }}) to personalize SQL queries based on user or context.
  • Security: Never expose raw URL parameters in queries without validation; use in controlled environments only.
  • Public Access: Enable via PUBLIC_ROLE_LIKE=gamma + assign dashboard permissions to Public role.
  • Geospatial Data: Use ISO codes (e.g., FR-01) to join data with built-in maps; simplify GeoJSON for performance.
  • Deck.gl: Ideal for point/heatmap visualizations; requires lat/long columns; supports Mapbox background with API key.

Tools Used

  • Apache Superset (v2+)
  • Google Sheets (for CSV data import)
  • QGIS (for GeoJSON simplification)
  • Mapbox (for custom basemaps, optional)
  • Docker (for Superset deployment)
  • Terminal / CLI (for feature flag and GeoJSON configuration)

Common Pitfalls

  • Inverted latitude/longitude fields causing data to render in wrong locations (e.g., Somalia instead of France).
  • Large GeoJSON files (>1MB) causing browser lag or failure to load.
  • CSV encoding or separator issues (commas vs. periods, UTF-8).
  • Missing PUBLIC_ROLE_LIKE flag preventing public access even with correct permissions.
  • Google Sheets not shared publicly → “Connection failed” error.

Practice Suggestions

  1. Create a Google Sheet with 5–10 rows of lat/long + metric (e.g., pollution levels, population).
  2. Connect to Superset and create a Deck.gl Scatterplot or Heatmap.
  3. Enable PUBLIC_ROLE_LIKE=gamma and share dashboard link with a non-logged-in user.
  4. Try adding a custom GeoJSON (e.g., French regions) and join with a dataset using FR-01 codes.
  5. Add a Markdown text block with icons or a title to a dashboard layout.