databricks-solutions · dilanpateldatabricks · May 26, 2026
diff --git a/demos/smart_factory/.gitignore b/demos/smart_factory/.gitignore
@@ -0,0 +1,11 @@
+__pycache__/
+*.pyc
+.env
+.venv/
+venv/
+node_modules/
+# frontend/dist/ — intentionally tracked for Databricks Apps deployment
+.databricks/
+*.egg-info/
+.DS_Store
+uv.lock
diff --git a/demos/smart_factory/Makefile b/demos/smart_factory/Makefile
@@ -0,0 +1,17 @@
+.PHONY: build-frontend deploy dev clean
+
+build-frontend:
+	cd frontend && npm ci && npm run build
+
+deploy: build-frontend
+	databricks bundle deploy -t dev
+
+deploy-prod: build-frontend
+	databricks bundle deploy -t prod
+
+dev:
+	cd frontend && npm run dev &
+	uvicorn app:app --reload --port 8000
+
+clean:
+	rm -rf frontend/dist frontend/node_modules
diff --git a/demos/smart_factory/README.md b/demos/smart_factory/README.md
@@ -0,0 +1,152 @@
+# SmartFactory — IoT Streaming Demo
+
+A customer-facing demo showing how Databricks turns factory sensor data into actionable insights — from machine floor to decision — with no Kafka, no infrastructure to manage, and full governance out of the box.
+
+## What This Demonstrates
+
+| Capability | Business Value |
+|---|---|
+| **ZeroBus Ingest** | Eliminate Kafka and message bus infrastructure. Sensor data flows directly into governed Delta tables. |
+| **SDP Streaming Pipeline** | Catch equipment anomalies as they happen, not in tomorrow's batch report. Continuous, serverless, pure SQL. |
+| **ML in the Pipeline** | Every sensor reading scored for anomalies inline — predictive maintenance without a separate ML platform. |
+| **Live Operations Dashboard** | Plant managers and technicians see machine health the moment it changes. Faster response, less downtime. |
+| **Unity Catalog Governance** | Every table governed from the first byte. Lineage, access control, audit — ready for compliance on day one. |
+| **Databricks Apps** | Full-stack app deployed and managed by Databricks. No separate hosting. |
+| **DABs** | Entire demo — app, pipeline, dashboard — deploys with a single command. |
+
+## Architecture
+
+![Architecture](docs/architecture.png)
+
+## The Machines
+
+| Machine | Sensors | Fault Scenario |
+|---|---|---|
+| **CNC Mill** | Temperature, Vibration, Spindle RPM | Overheating, vibration spike |
+| **Hydraulic Press** | Pressure, Temperature, Cycle Count | Pressure surge, cycle slowdown |
+| **Conveyor Belt** | Belt Speed, Load Weight, Motor Current | Speed drop, overcurrent |
+
+## Demo Story (6 minutes)
+
+> **Pre-flight**: Start streaming + pipeline 30s before presenting. Confirm dashboard has data.
+
+### Act 1 — "This is your factory" (IoT Simulation tab)
+> "3 machines, IoT sensors streaming every 2 seconds, directly into Databricks. No Kafka."
+
+- Gauges are already updating, event feed scrolling
+- Point out "Streaming" and "Pipeline Running" in the header
+- Expand ZeroBus info panel — highlight ≤200ms ack, 10 GB/s, Joby Aviation quote
+
+### Act 2 — "Here's the pipeline" (SDP in Databricks UI)
+> "Declarative SQL. Streaming and batch in one pipeline. Fully serverless."
+
+- Switch to Databricks workspace, open the SDP pipeline DAG
+- Show Bronze → Silver → Gold with streaming indicators
+- Click into Silver SQL — "anomaly detection is a SQL JOIN. Any SQL developer can own this."
+- Three SDP benefits: declarative, streaming+batch unified, serverless
+
+### Act 3 — "Let's break something" (Inject a fault)
+> "Watch what happens when the CNC Mill starts overheating."
+
+- Click **Fault: CNC Mill** — watch temperature climb, gauges go red
+- Event feed lights up with warnings and criticals
+- Switch to **Operations Dashboard** — health scores dropping, anomaly log filling
+
+### Act 4 — "Everything is governed" (Unity Catalog)
+> "Every table governed. Full lineage from raw sensor event to dashboard."
+
+- Open Catalog Explorer, click Gold table → show lineage graph
+- "One command to deploy. No Kafka. No ML infrastructure. Just push and go."
+
+### Act 5 — "Clear the fault" (Resolution)
+- Click **Clear All** — readings normalize, health scores recover
+
+See [docs/demo-script.md](docs/demo-script.md) for the full script with talking points and objection handling.
+
+## Quick Start
+
+### Prerequisites
+- [Databricks CLI](https://docs.databricks.com/dev-tools/cli/index.html) v0.288+ installed
+- CLI profile configured for your target workspace
+- Node.js 18+ and npm installed
+- An existing Unity Catalog catalog with cloud storage configured
+
+### One-command setup
+```bash
+git clone <repo-url>
+cd smartfactory-demo
+./setup.sh <databricks-cli-profile> <catalog_name>
+```
+
+This script handles everything:
+1. Finds and starts a SQL warehouse
+2. Creates the `smartfactory` schema and landing table
+3. Builds the React frontend
+4. Deploys all resources via DABs (app, pipeline, dashboard)
+5. Detects the app service principal and grants all permissions
+6. Starts the app and deploys code
+7. Sets the SDP pipeline to continuous mode
+
+### After setup
+1. Open the app URL printed by the script
+2. Click **Streaming** in the header to start the data simulator
+3. Click **Start Pipeline** to begin continuous SDP processing
+4. Switch between **IoT Simulation** and **Operations Dashboard** tabs
+5. Inject faults and watch anomalies flow through the pipeline
+
+> **Important: When you're done demoing, stop the streaming and pipeline!**
+> Both consume compute resources. Click **Streaming** (to pause) and **Pipeline Running** (to stop) in the header bar.
+> The simulator and pipeline start paused by default — you must manually start them each demo session.
+
+### Redeploying after code changes
+```bash
+cd frontend && npm run build && cd ..
+databricks bundle deploy -t dev
+databricks apps deploy smartfactory-app \
+  --source-code-path /Workspace/Users/<you>/.bundle/smartfactory-demo/dev/files
+```
+
+## Project Structure
+
+```
+smartfactory-demo/
+├── setup.sh                  # One-command setup script
+├── databricks.yml            # DABs bundle (app + pipeline + dashboard)
+├── app.yaml                  # Databricks App config
+├── app.py                    # FastAPI backend (WebSocket + REST + pipeline control)
+├── simulator.py              # 3-machine IoT sensor simulator with fault injection
+├── zerobus_client.py         # ZeroBus SDK wrapper with SQL INSERT fallback
+├── pipeline/
+│   ├── bronze.sql            # Validated ingestion (streaming table)
+│   ├── silver.sql            # Anomaly scoring via threshold JOIN (streaming table)
+│   └── gold.sql              # Health KPIs + anomaly timeline (materialized views)
+├── frontend/
+│   ├── src/
+│   │   ├── App.tsx           # Tabbed layout (IoT Simulation + Dashboard)
+│   │   ├── components/
+│   │   │   ├── FactoryFloor  # SVG machine visuals with live sensor readouts
+│   │   │   ├── MachineCard   # Per-machine gauge cards
+│   │   │   ├── SensorGauge   # Circular SVG gauge component
+│   │   │   ├── ControlPanel  # Fault injection buttons
+│   │   │   ├── EventFeed     # Live scrolling event log
+│   │   │   ├── DashboardView # Charts, KPI tables, anomaly log
+│   │   │   └── PipelineBanner# Pipeline flow + UC governance badge
+│   │   └── hooks/
+│   │       └── useWebSocket  # Auto-reconnecting WebSocket hook
+│   └── dist/                 # Pre-built frontend (deployed with app)
+├── dashboard.lvdash.json     # Lakeview dashboard definition
+└── CLAUDE.md                 # Development notes and known issues
+```
+
+## Tech Stack
+
+| Layer | Technology |
+|---|---|
+| Frontend | React 18, TypeScript, Vite, TailwindCSS, Recharts |
+| Backend | FastAPI, Uvicorn, WebSocket |
+| Ingestion | ZeroBus SDK (with SQL INSERT fallback) |
+| Pipeline | SDP (Spark Declarative Pipelines), serverless |
+| ML | SQL threshold-based anomaly detection in SDP Silver layer |
+| Governance | Unity Catalog (lineage, access control, audit) |
+| Deployment | Databricks Asset Bundles (DABs) |
+| Dashboard | Lakeview (AI/BI) + in-app React dashboard |
diff --git a/demos/smart_factory/app.yaml b/demos/smart_factory/app.yaml
@@ -0,0 +1,19 @@
+command: ["uvicorn", "src.app:app", "--host", "0.0.0.0", "--port", "8000"]
+
+env:
+  - name: CATALOG_NAME
+    value: dilan_catalog
+  - name: LANDING_SCHEMA
+    value: smartfactory
+  - name: ZEROBUS_TABLE
+    value: dilan_catalog.smartfactory.raw_sensor_events
+  - name: WAREHOUSE_ID
+    value: "01c7fe04f060528e"
+  - name: PIPELINE_SCHEMA
+    value: dev_dilan_patel_smartfactory
+  - name: PIPELINE_ID
+    value: "4b993ed3-336f-40b5-8bb7-55ff9f056707"
+  - name: ENABLE_SIMULATOR
+    value: "false"
+  - name: SIMULATOR_INTERVAL_MS
+    value: "1000"
diff --git a/demos/smart_factory/dashboards/smartfactory.lvdash.json b/demos/smart_factory/dashboards/smartfactory.lvdash.json
@@ -0,0 +1,145 @@
+{
+  "pages": [
+    {
+      "name": "machine-health-overview",
+      "displayName": "Machine Health Overview",
+      "layout": [
+        {
+          "widget": {
+            "name": "machine_health_scores",
+            "queries": [
+              {
+                "name": "main_query",
+                "query": {
+                  "datasetName": "machine_summary",
+                  "disaggregated": false
+                }
+              }
+            ],
+            "spec": {
+              "version": 2,
+              "widgetType": "bar",
+              "encodings": {
+                "x": { "fieldName": "machine_id", "displayName": "Machine" },
+                "y": { "fieldName": "avg_health_score", "displayName": "Health Score" },
+                "color": { "fieldName": "machine_type", "displayName": "Type" }
+              }
+            }
+          },
+          "position": { "x": 0, "y": 0, "width": 3, "height": 2 }
+        },
+        {
+          "widget": {
+            "name": "anomaly_counts",
+            "queries": [
+              {
+                "name": "main_query",
+                "query": {
+                  "datasetName": "machine_summary",
+                  "disaggregated": false
+                }
+              }
+            ],
+            "spec": {
+              "version": 2,
+              "widgetType": "bar",
+              "encodings": {
+                "x": { "fieldName": "machine_id", "displayName": "Machine" },
+                "y": { "fieldName": "total_criticals", "displayName": "Critical Alerts" },
+                "color": { "fieldName": "machine_id", "displayName": "Machine" }
+              }
+            }
+          },
+          "position": { "x": 3, "y": 0, "width": 3, "height": 2 }
+        }
+      ]
+    },
+    {
+      "name": "sensor-trends",
+      "displayName": "Sensor Trends",
+      "layout": [
+        {
+          "widget": {
+            "name": "sensor_timeseries",
+            "queries": [
+              {
+                "name": "main_query",
+                "query": {
+                  "datasetName": "enriched_events",
+                  "disaggregated": false
+                }
+              }
+            ],
+            "spec": {
+              "version": 2,
+              "widgetType": "line",
+              "encodings": {
+                "x": { "fieldName": "timestamp", "displayName": "Time" },
+                "y": { "fieldName": "value", "displayName": "Sensor Value" },
+                "color": { "fieldName": "sensor_name", "displayName": "Sensor" }
+              }
+            }
+          },
+          "position": { "x": 0, "y": 0, "width": 6, "height": 3 }
+        }
+      ]
+    },
+    {
+      "name": "anomaly-log",
+      "displayName": "Anomaly Log",
+      "layout": [
+        {
+          "widget": {
+            "name": "anomaly_table",
+            "queries": [
+              {
+                "name": "main_query",
+                "query": {
+                  "datasetName": "anomaly_timeline",
+                  "disaggregated": false
+                }
+              }
+            ],
+            "spec": {
+              "version": 2,
+              "widgetType": "table",
+              "encodings": {
+                "columns": [
+                  { "fieldName": "timestamp", "displayName": "Time" },
+                  { "fieldName": "machine_id", "displayName": "Machine" },
+                  { "fieldName": "sensor_name", "displayName": "Sensor" },
+                  { "fieldName": "value", "displayName": "Value" },
+                  { "fieldName": "anomaly_status", "displayName": "Status" },
+                  { "fieldName": "critical_threshold", "displayName": "Threshold" }
+                ]
+              }
+            }
+          },
+          "position": { "x": 0, "y": 0, "width": 6, "height": 4 }
+        }
+      ]
+    }
+  ],
+  "datasets": [
+    {
+      "name": "machine_summary",
+      "displayName": "Machine Summary",
+      "query": "SELECT * FROM dilan_catalog.smartfactory.gold.machine_summary"
+    },
+    {
+      "name": "enriched_events",
+      "displayName": "Enriched Events",
+      "query": "SELECT * FROM dilan_catalog.smartfactory.silver.enriched_events WHERE timestamp > current_timestamp() - INTERVAL 1 HOUR ORDER BY timestamp DESC LIMIT 5000"
+    },
+    {
+      "name": "anomaly_timeline",
+      "displayName": "Anomaly Timeline",
+      "query": "SELECT * FROM dilan_catalog.smartfactory.gold.anomaly_timeline LIMIT 500"
+    },
+    {
+      "name": "health_kpis",
+      "displayName": "Health KPIs",
+      "query": "SELECT * FROM dilan_catalog.smartfactory.gold.machine_health_kpis"
+    }
+  ]
+}