Superset x Claude MCP — Meeting Room Utilisation

AI-automated meeting room analytics: 50,000+ bookings across 8 floors, 50+ rooms, with heatmaps, utilisation tracking, and native filters — built entirely through Claude MCP.

---

What This Is

A comprehensive meeting room utilisation dashboard replicating a real enterprise corporate analytics report. Every chart, filter, and layout was created programmatically via Apache Superset's API — no manual UI interaction.

Architecture

┌─────────────────────────┐
│   PostgreSQL 16         │  50,273 synthetic booking records
│   (meeting_bookings)    │  Single fact table + denormalised dims
└────────┬────────────────┘
         │ SQLAlchemy
┌────────▼────────────────┐
│   Apache Superset 5.0+  │  Dashboard engine
│   (Docker Compose)      │  13 charts, 6 native filters
└────────┬────────────────┘
         │ MCP Server + REST API
┌────────▼────────────────┐
│   Claude Code / MCP     │  Natural language → API calls
└─────────────────────────┘

Key Metrics

Metric	Value
Total Bookings	50,273
Total Hours Booked	~43,092
Avg Booking Duration	~51 minutes
Utilisation Rate	~27.27%
Date Range	Jan 2024 — Dec 2026
Rooms	48 across 8 floors

13 Charts

#	Chart	Type
1	Total Bookings	KPI Card
2	Total Hours Booked	KPI Card
3	Average Booking Duration	KPI Card
4	Utilisation Rate	KPI Card
5	Bookings by Floor Level	Donut
6	Bookings by Time	Donut
7	Bookings by Day of Week	Donut
8	Booking Heat Map (Peak Times)	Heatmap
9	Bookings by Floor Level	Horizontal Bar
10	Bookings & Duration by Day of Week	Combo (Bar + Line)
11	Bookings by Room	Horizontal Bar
12	Utilisation Rate by Room	Horizontal Bar
13	Bookings by Date	Time Series Area

6 Native Filters

Date Range, Hour, Organiser Name, Subject, Room, Floor Level — with cross-filtering enabled.

Colour Palette

Role	HEX	Usage
Primary	#002664	KPI text, time series, line overlays
Secondary	#146cfd	Emphasis, secondary charts
Teal	#2e808e	Category splits
Baby Blue	#8ce0ff	Backgrounds
Soft Pink	#ffb8c1	Donut segments, bar charts
Alert	#630019	Heatmap high, utilisation bars
Neutral	#d1eeea	Heatmap low, gridlines

Data Model

Single fact table with denormalised dimensions for simplicity:

meeting_bookings (50,273 rows)
├── booking_id, organizer_name, room_name, floor_level, floor_level_name
├── subject, start_datetime, end_datetime, start_date
├── day_of_week, day_of_week_number, hour_ampm, hour_of_day
├── duration_minutes, is_cancelled
└── total_available_minutes, total_minutes_booked, utilisation_rate

48 rooms across 8 floors (Level 2–9), naming: {floor}.{wing}.{seq}-S{size}-VC

200 organisers across 6 dummy regions (Region A–E + Central Office)

Quick Start

git clone https://github.com/yujiyamane/superset-claude-meeting.git
cd superset-claude-meeting

# Start Superset + PostgreSQL
docker compose up -d

# Load data and build dashboard
python scripts/load_meeting_data.py
python scripts/create_meeting_charts.py
python scripts/create_meeting_dashboard.py

# Open http://localhost:8088

Testing

28 tests across 3 test files:

python -m pytest tests/ -v

Test File	Tests	Coverage
test_meeting_data.py	16	Data quality, format, org name safety
test_meeting_charts.py	9	Chart existence, types, counts
test_meeting_dashboard.py	3	Dashboard, chart links, filters

Repository Structure

superset-claude-meeting/
├── README.md
├── docker-compose.yml
├── requirements.txt
├── superset/
│   ├── color_theme.py
│   ├── superset_config.py
│   └── dashboards/
│       └── meeting_room_utilisation.zip
├── data/
│   ├── generate_meeting_data.py
│   └── meeting_schema.sql
├── scripts/
│   ├── superset_client.py
│   ├── load_meeting_data.py
│   ├── create_meeting_charts.py
│   ├── create_meeting_dashboard.py
│   ├── export_meeting_dashboard.py
│   └── take_meeting_screenshots.py
├── tests/
│   ├── test_meeting_data.py
│   ├── test_meeting_charts.py
│   └── test_meeting_dashboard.py
└── screenshots/
    ├── meeting_dashboard_overview.png
    ├── meeting_kpi_cards.png
    ├── meeting_donuts.png
    └── meeting_heatmap.png

Related Projects

Case	Focus	Repo
Case 1	Power BI x Healthcare KPI	powerbi-claude-health
Case 2	Streamlit x ED Performance	streamlit-claude-health
Case 3	Life OS x Personal Analytics	lifeos-claude-dashboard
Case 4	Superset x Healthcare ED + RLS	superset-claude-health
Case 5	Superset x Meeting Room Utilisation	This repo

Author

Yuji Yamane — Senior Engineer / AI-Augmented BI Solutions Lead