quicksight-codegen

Generate and deploy Amazon QuickSight dashboards from any CSV file. One command, zero manual configuration.

# Preview locally (no AWS needed)
quicksight-codegen preview --csv your_data.csv

# Deploy to QuickSight
quicksight-codegen deploy --csv your_data.csv --dataset "your-dataset" --name "My Dashboard"

The tool automatically analyzes your CSV columns, picks the right chart types (KPI, bar, pie, heatmap, table), adds interactive filters, and handles data type conversion — all as code.

Quick Start

1. Install

git clone https://github.com/AzerMusic/esg-quicksight-as-code.git
cd esg-quicksight-as-code

# Using uv (recommended)
uv venv && uv pip install -e ".[aws,auto]"

# Or using pip
python -m venv .venv && source .venv/bin/activate
pip install -e ".[aws,auto]"

2. Preview locally (no AWS required)

quicksight-codegen preview --csv your_data.csv --name "My Dashboard"

This generates an HTML file with interactive Chart.js charts you can open in any browser.

3. Deploy to AWS QuickSight

Prerequisites:

• AWS CLI configured (aws configure)
• QuickSight activated in your AWS account
• A dataset uploaded via QuickSight web console ("Upload a file")

Setup environment:

cp .env.example .env
# Edit .env with your values (see "Finding Your AWS Values" below)

Deploy:

# List your datasets to find the right name
quicksight-codegen list-datasets

# Deploy
quicksight-codegen deploy \
  --csv your_data.csv \
  --dataset "your-dataset-name" \
  --name "My Dashboard"

The tool will:

• Auto-detect your AWS account ID, region, and QuickSight user
• Generate KPIs, charts, heatmaps, and tables based on your data
• Add dropdown filters for categorical columns (Sector, Region, etc.)
• Create parseDecimal() calculated fields so numeric columns work without manual type fixing
• Deploy as a QuickSight Analysis you can open immediately

Finding Your AWS Values

Most of these are auto-detected if your AWS CLI is configured. You only need .env if auto-detection doesn't work for your setup.

CLI Commands

Deploy options

quicksight-codegen deploy \
  --csv data.csv \            # Path to CSV or Excel file
  --dataset "my-dataset" \    # Dataset name in QuickSight
  --name "My Dashboard" \     # Dashboard title
  --id "my-dashboard-v1" \    # Analysis ID (optional, auto-generated from name)
  --region eu-central-1 \     # AWS region (optional, auto-detected)
  --update \                  # Update existing analysis instead of creating new
  --dry-run \                 # Generate JSON only, don't deploy
  --fix-types                 # Fix column types before deploying (SPICE datasets only)

How It Works

CSV file
  |
  v
Column analysis (numeric vs categorical)
  |
  +--> KPI visuals for key numeric columns
  +--> Bar/Pie charts for numeric × categorical
  +--> Heatmap for numeric × 2 categoricals
  +--> Data table with all columns
  +--> Dropdown filters for categorical columns
  +--> parseDecimal() calculated fields for numeric columns
  |
  v
QuickSight Analysis JSON + HTML preview
  |
  v
Deploy via boto3 API

Key feature: automatic type conversion. When you upload a CSV to QuickSight, all columns become STRING type. Charts need DECIMAL columns for numeric data. This tool automatically generates parseDecimal() calculated fields in the Analysis definition, so charts work immediately without manual type fixing in the QuickSight console.

Python API

For programmatic use, the library exposes all QuickSight visual types:

from quicksight_codegen import (
    auto_dashboard,          # Auto-generate from CSV
    BarChartVisual,          # Manual visual construction
    KPIVisual,
    PieChartVisual,
    TableVisual,
    HeatMapVisual,
    deploy_analysis,         # Deploy to AWS
    simulate_deploy,         # Generate JSON without AWS
    generate_html_preview,   # Local HTML preview
)

# Auto mode: one function call
analysis, html_path = auto_dashboard(
    data="your_data.csv",
    name="My Dashboard",
    output_dir="output/",
)

# Manual mode: full control over each visual
bar = BarChartVisual("bar-1")
bar.add_categorical_dimension_field("Region", "dataset")
bar.add_numerical_measure_field("Revenue", "dataset", "SUM")

See examples/ for complete examples.

Supported Visual Types

Basic: BarChart, LineChart, Table, PivotTable, KPI, PieChart, ScatterPlot

Advanced: TreeMap, HeatMap, FilledMap, GeospatialMap, FunnelChart, Waterfall, BoxPlot, GaugeChart

Project Structure

src/quicksight_codegen/
  __init__.py         # Public API exports
  auto.py             # Auto dashboard generation from CSV
  cli.py              # CLI tool (preview, deploy, list-datasets, fix-types)
  analysis.py         # Analysis definition, sheets, calculated fields
  deploy.py           # AWS QuickSight deployment via boto3
  discovery.py        # Auto-detect AWS account, region, user, datasets
  preview.py          # HTML preview with Chart.js
  filters.py          # Filter types and dropdown controls
  dataset.py          # Dataset type inference and fixing
  visuals/            # All visual type classes
  templates/          # Pre-built dashboard templates (ESG, portfolio)

Running Tests

uv pip install -e ".[dev]"
uv run pytest

Authors

• Weihua SHI
• Yasmine ZEROUAL
• Adam TOUCHANE
• Saber BERREHILI
• Amine MEGHAGHI

License

MIT