Exploring Apache Superset 2.0: From Data Exploration to Modern BI Dashboards

Introduction – Zhao Yongjie, a Superset PMC member and Preset senior data engineer, presents "From Exploratory Data Analysis to Modern BI Dashboards: Superset 2.0".

1. Apache Superset History – Originated from an Airbnb hackathon in 2015, matured over seven years, and became an Apache top‑level project in early 2022; it now enjoys broad community support from companies like Apple and Dropbox.

2. Core Features

• Superset is a Python‑based project that uses the DBAPI2 interface to connect to any relational database, allowing easy addition of new dialects. • SQLLab IDE offers a powerful, Hue‑like SQL editor whose visual datasets feed charts and dashboards. • Over 50 built‑in visualizations can be extended via a plugin system. • Charts can be placed on dashboards with rich filtering, CSS templating, row‑level security, and scheduled reports. • Native filters, cross‑filtering, drill‑through, auto‑refresh, and sharing options (URL, email, iframe) enhance interactivity.

3. Data Exploration – Drag‑and‑drop UI lets users combine metrics and dimensions, switch chart types instantly, customize SQL snippets, preview results and samples, perform advanced time analysis (shifts, rolling calculations, forecasting), and add annotation layers.

4. Demo of Data Exploration – Live walkthrough using the Superstore dataset shows metric/dimension selection, time‑grain adjustments, chart type changes, time‑shift queries, and SQL preview/copy functionality.

5. Dashboards – Dashboards aggregate charts, support native filters, multi‑tab layouts, markdown, cross‑filtering, drill‑through to detail, chart maximization, and automatic refresh. Demonstrations include native filter configuration, filter scope control, and SQL inspection.

6. Tips

1) Feature Flags – Experimental features are disabled by default; enable them in config.py (e.g., CLIENT_CACHE, DRILL_TO_DETAIL). 2) Custom Visualization Plugins – Superset provides a front‑end scaffolding tool to create custom viz plugins with minimal code; a tutorial is linked for quick implementation.

7. Q&A

Q1: Superset’s query API can be consumed by external applications. Q2: Queries run directly against the underlying DB/warehouse; performance depends on the downstream system. Q3: Distinct‑based filter dropdowns can be optimized by pointing to dimension tables, though the open‑source version requires custom hacks.

8. Community & Resources – Join the Superset GitHub repo, Slack channel, Apache mailing list, and Preset blog for issues, PRs, discussions, and tutorials. Links to the data‑intelligence knowledge map and business‑cooperation contacts are also provided.