Creating Dashboards

Purpose

This skill enables the creation of sophisticated dashboard interfaces that aggregate and present data through coordinated widgets including KPI cards, charts, tables, and filters. Dashboards serve as centralized command centers for data-driven decision making, combining multiple component types from other skills (data-viz, tables, design-tokens) into unified analytics experiences with real-time updates, responsive layouts, and interactive filtering.

When to Use

Activate this skill when:

• Building business intelligence or analytics dashboards
• Creating executive reporting interfaces
• Implementing real-time monitoring systems
• Designing KPI displays with metrics and trends
• Developing customizable widget-based layouts
• Coordinating filters across multiple data displays
• Building responsive data-heavy interfaces
• Implementing drag-and-drop dashboard editors
• Creating template-based analytics systems
• Designing multi-tenant SaaS dashboards

Core Dashboard Elements

KPI Card Anatomy

┌────────────────────────────┐
│ Revenue (This Month)       │ ← Label with time period
│                            │
│  $1,245,832               │ ← Big number (primary metric)
│  ↑ 15.3% vs last month    │ ← Trend indicator with comparison
│  ▂▃▅▆▇█ (sparkline)       │ ← Mini visualization
└────────────────────────────┘

Widget Container Structure

• Title bar with widget name and actions
• Loading state (skeleton or spinner)
• Error boundary with retry option
• Resize handles for adjustable layouts
• Settings menu (export, configure, refresh)

Dashboard Layout Types

Fixed Layout: Designer-defined placement, consistent across users Customizable Grid: User drag-and-drop, resizable widgets, saved layouts Template-Based: Pre-built patterns, industry-specific starting points

Global Dashboard Controls

• Date range picker (affects all widgets)
• Filter panel (coordinated across widgets)
• Refresh controls (manual/auto-refresh)
• Export actions (PDF, image, data)
• Theme switcher (light/dark/custom)

Implementation Approach

1. Choose Dashboard Architecture

For Quick Analytics Dashboard → Use Tremor Pre-built KPI cards, charts, and tables with minimal code:

npm install @tremor/react

For Customizable Dashboard → Use react-grid-layout Drag-and-drop, resizable widgets, user-defined layouts:

npm install react-grid-layout

2. Set Up Global State Management

Implement filter context for cross-widget coordination:

// Dashboard context for shared filters
const DashboardContext = createContext({
  filters: { dateRange: null, categories: [] },
  setFilters: () => {},
  refreshInterval: 30000
});

// Wrap dashboard with provider
<DashboardContext.Provider value={dashboardState}>
  <FilterPanel />
  <WidgetGrid />
</DashboardContext.Provider>

3. Implement Data Fetching Strategy

Parallel Loading: Fetch all widget data simultaneously Lazy Loading: Load visible widgets first, others on scroll Cached Updates: Serve from cache while fetching fresh data

4. Configure Real-Time Updates

Server-Sent Events (Recommended for Dashboards):

const eventSource = new EventSource('/api/dashboard/stream');
eventSource.onmessage = (event) => {
  const update = JSON.parse(event.data);
  updateWidget(update.widgetId, update.data);
};

5. Apply Responsive Design

Define breakpoints for different screen sizes:

• Desktop (>1200px): Multi-column grid
• Tablet (768-1200px): 2-column layout
• Mobile (<768px): Single column stack

Quick Start with Tremor

Basic KPI Dashboard

import { Card, Grid, Metric, Text, BadgeDelta, AreaChart } from '@tremor/react';

function QuickDashboard({ data }) {
  return (
    <Grid numItems={1} numItemsSm={2} numItemsLg={4} className="gap-4">
      {/* KPI Cards */}
      <Card>
        <Text>Total Revenue</Text>
        <Metric>$45,231.89</Metric>
        <BadgeDelta deltaType="increase">+12.5%</BadgeDelta>
      </Card>

      <Card>
        <Text>Active Users</Text>
        <Metric>1,234</Metric>
        <BadgeDelta deltaType="decrease">-2.3%</BadgeDelta>
      </Card>

      {/* Chart Widget */}
      <Card className="lg:col-span-2">
        <Text>Revenue Trend</Text>
        <AreaChart
          data={data.revenue}
          index="date"
          categories={["revenue"]}
          valueFormatter={(value) => `$${value.toLocaleString()}`}
        />
      </Card>
    </Grid>
  );
}

For complete implementation, see examples/tremor-dashboard.tsx.

Customizable Dashboard Implementation

Drag-and-Drop Grid Layout

import { Responsive, WidthProvider } from 'react-grid-layout';
import 'react-grid-layout/css/styles.css';

const ResponsiveGridLayout = WidthProvider(Responsive);

function CustomizableDashboard() {
  const [layouts, setLayouts] = useState(getStoredLayouts());

  return (
    <ResponsiveGridLayout
      layouts={layouts}
      breakpoints={{ lg: 1200, md: 996, sm: 768 }}
      cols={{ lg: 12, md: 10, sm: 6 }}
      rowHeight={60}
      onLayoutChange={(layout, layouts) => {
        setLayouts(layouts);
        localStorage.setItem('dashboardLayout', JSON.stringify(layouts));
      }}
      draggableHandle=".widget-header"
    >
      <div key="kpi1">
        <KPIWidget data={kpiData} />
      </div>
      <div key="chart1">
        <ChartWidget data={chartData} />
      </div>
      <div key="table1">
        <TableWidget data={tableData} />
      </div>
    </ResponsiveGridLayout>
  );
}

For full example with widget catalog, see examples/customizable-dashboard.tsx.

Real-Time Data Patterns

Server-Sent Events (Recommended)

Best for unidirectional updates from server to dashboard:

function useSSEUpdates(endpoint) {
  useEffect(() => {
    const eventSource = new EventSource(endpoint);

    eventSource.onmessage = (event) => {
      const update = JSON.parse(event.data);
      // Update specific widget or all widgets
      dispatch({ type: 'UPDATE_WIDGET', payload: update });
    };

    return () => eventSource.close();
  }, [endpoint]);
}

WebSocket (For Bidirectional)

Use when dashboard needs to send commands back to server:

const ws = new WebSocket('ws://localhost:3000/dashboard');
ws.onmessage = (event) => {
  const data = JSON.parse(event.data);
  updateDashboard(data);
};
// Send filter changes to server
ws.send(JSON.stringify({ type: 'FILTER_CHANGE', filters }));

Smart Polling Fallback

For environments without WebSocket/SSE support:

function useSmartPolling(fetchData, interval = 30000) {
  const [isPaused, setIsPaused] = useState(false);

  useEffect(() => {
    if (isPaused || document.hidden) return;

    const timer = setInterval(fetchData, interval);
    return () => clearInterval(timer);
  }, [isPaused, interval]);

  // Pause when tab inactive
  useEffect(() => {
    const handleVisibilityChange = () => {
      setIsPaused(document.hidden);
    };
    document.addEventListener('visibilitychange', handleVisibilityChange);
    return () => document.removeEventListener('visibilitychange', handleVisibilityChange);
  }, []);
}

For detailed patterns including error handling and reconnection, see references/real-time-updates.md.

Performance Optimization

Lazy Loading Strategy

function DashboardGrid({ widgets }) {
  const [visibleWidgets, setVisibleWidgets] = useState(new Set());

  return widgets.map(widget => (
    <LazyLoad
      key={widget.id}
      height={widget.height}
      offset={100}
      once
      placeholder={<WidgetSkeleton />}
    >
      <Widget {...widget} />
    </LazyLoad>
  ));
}

Parallel Data Fetching

// Fetch all widget data simultaneously
const loadDashboard = async () => {
  const [kpis, charts, tables] = await Promise.all([
    fetchKPIs(),
    fetchChartData(),
    fetchTableData()
  ]);

  return { kpis, charts, tables };
};

Widget