AMG Organization Dashboard

Zad Rafi

AMG Organization Dashboard

Real-time store estate visualization from Snowflake

Author

Affiliation

Zad Rafi

Less Likely

Published

June 22, 2026

This page documents the AMG Organization Dashboard, a real-time interactive visualization of the AMG store estate (346 locations) powered by Snowflake and D3.

Overview

The dashboard provides a comprehensive view of AMG’s organization across divisions, regions, districts, and class-of-trade segments. Data is fetched live from the Snowflake MARKETBASKET.PUBLIC.ORGANIZATION table.

Status: Separate service running on http://localhost:8540

Getting Started

Installation

Navigate to the dashboard directory and install dependencies:

cd my-quarto-site/dashboards/amg-organization
pip install -r requirements.txt

Start the Server

python3 amg-dashboard/server.py

You should see:

 * Running on http://127.0.0.1:8540
 * Press CTRL+C to quit

Access the Dashboard

Open http://localhost:8540 in your browser.

Features

Live Data

✓ LIVE DATA indicator shows when Snowflake is reachable
Real-time sync on page load from /api/organization
Falls back to sample data if Snowflake is unreachable

Interactive Filters

Division: Company Ops, Dealers
Region: NYC Metro, Dealer, others
COT (Class of Trade): Company Op, Supply Only, Liquor Store, Closed
Fuel Brand: BP, Mobil, Shell, Citgo, Gulf, Exxon, Sunoco, Unbranded
SSS YTD & SSS Rolling: Same-store-sales performance categories

Click any bar or donut segment to filter; click again to clear.

Visualizations

Chart	Insight
KPIs	Total sites, open count, company ops, dealers, closed locations
Sites by Region	Bar chart of store count per region (interactive)
COT	Donut breakdown by class of trade (Company Op, Supply Only, Liquor, Closed)
Fuel Brand	Donut by fuel brand distribution
Division	Pie of Company Ops vs. Dealers
State	Pie of store count by state (NY, NJ, CT)
Geographic	Scatter plot of store locations by GPS coordinates
Hierarchy	Treemap of Division → Region → District
Site Detail	Sortable table of all 346 sites with tagging

Export & Import

Export CSV: Download filtered data as CSV
Import JSON: Replace live data with a local .json export for offline analysis

Configuration

Snowflake Connection

The server reads credentials from ~/.snowflake/connections.toml:

[connections]
[connections.ZADADMIN]
account = "your_account"
user = "your_user"
password = "your_password"  # or use SSO
warehouse = "COMPUTE_WH"

To use a different connection name, edit CONNECTION_NAME in server.py.

Data Query

To modify the columns or table:

Edit the QUERY string in server.py
Update the COLS array in index.html to match
Restart the server and refresh

Directory Structure

my-quarto-site/dashboards/amg-organization/
├── index.html          # D3 dashboard (single-page app)
├── server.py           # Flask backend + Snowflake connector
├── requirements.txt    # Python dependencies
└── README.md           # Detailed documentation

Troubleshooting

Issue	Solution
“SAMPLE — Snowflake unreachable”	Check VPN, credentials in `~/.snowflake/connections.toml`, and warehouse access
ModuleNotFoundError	Run `pip install -r requirements.txt`
Port 8540 in use	Edit `app.rn(host="127.0.0.1", port=9999)` in `server.py`
No data appears	Verify query and Snowflake table permissions

Offline Mode

If Snowflake is temporarily unavailable:

Click ↓ import .json in the dashboard
Select a previously exported .json file
The dashboard will show local data while Snowflake is down

Next Steps

Monitoring: Set up health checks for Snowflake availability
Caching: Add Redis caching layer for high-traffic periods
Auth: Integrate Okta or Entra ID for production deployment
Deployment: Deploy to Azure Container Apps or Netlify Functions

--- title: "AMG Organization Dashboard" subtitle: "Real-time store estate visualization from Snowflake" date: today --- This page documents the **AMG Organization Dashboard**, a real-time interactive visualization of the AMG store estate (346 locations) powered by Snowflake and D3. ## Overview The dashboard provides a comprehensive view of AMG's organization across divisions, regions, districts, and class-of-trade segments. Data is fetched live from the Snowflake `MARKETBASKET.PUBLIC.ORGANIZATION` table. **Status**: Separate service running on `http://localhost:8540` ## Getting Started ### Installation Navigate to the dashboard directory and install dependencies: ```bash cd my-quarto-site/dashboards/amg-organization pip install -r requirements.txt ``` ### Start the Server ```bash python3 amg-dashboard/server.py ``` You should see: ``` * Running on http://127.0.0.1:8540 * Press CTRL+C to quit ``` ### Access the Dashboard Open **[http://localhost:8540](http://localhost:8540)** in your browser. ## Features ### Live Data - ✓ **LIVE DATA** indicator shows when Snowflake is reachable - Real-time sync on page load from `/api/organization` - Falls back to sample data if Snowflake is unreachable ### Interactive Filters - **Division**: Company Ops, Dealers - **Region**: NYC Metro, Dealer, others - **COT** (Class of Trade): Company Op, Supply Only, Liquor Store, Closed - **Fuel Brand**: BP, Mobil, Shell, Citgo, Gulf, Exxon, Sunoco, Unbranded - **SSS YTD** & **SSS Rolling**: Same-store-sales performance categories Click any bar or donut segment to filter; click again to clear. ### Visualizations | Chart | Insight | |-------|---------| | **KPIs** | Total sites, open count, company ops, dealers, closed locations | | **Sites by Region** | Bar chart of store count per region (interactive) | | **COT** | Donut breakdown by class of trade (Company Op, Supply Only, Liquor, Closed) | | **Fuel Brand** | Donut by fuel brand distribution | | **Division** | Pie of Company Ops vs. Dealers | | **State** | Pie of store count by state (NY, NJ, CT) | | **Geographic** | Scatter plot of store locations by GPS coordinates | | **Hierarchy** | Treemap of Division → Region → District | | **Site Detail** | Sortable table of all 346 sites with tagging | ### Export & Import - **Export CSV**: Download filtered data as CSV - **Import JSON**: Replace live data with a local `.json` export for offline analysis ## Configuration ### Snowflake Connection The server reads credentials from `~/.snowflake/connections.toml`: ```toml [connections] [connections.ZADADMIN] account = "your_account" user = "your_user" password = "your_password" # or use SSO warehouse = "COMPUTE_WH" ``` To use a different connection name, edit `CONNECTION_NAME` in `server.py`. ### Data Query To modify the columns or table: 1. Edit the `QUERY` string in `server.py` 2. Update the `COLS` array in `index.html` to match 3. Restart the server and refresh ## Directory Structure ``` my-quarto-site/dashboards/amg-organization/ ├── index.html # D3 dashboard (single-page app) ├── server.py # Flask backend + Snowflake connector ├── requirements.txt # Python dependencies └── README.md # Detailed documentation ``` ## Troubleshooting | Issue | Solution | |-------|----------| | "SAMPLE — Snowflake unreachable" | Check VPN, credentials in `~/.snowflake/connections.toml`, and warehouse access | | ModuleNotFoundError | Run `pip install -r requirements.txt` | | Port 8540 in use | Edit `app.rn(host="127.0.0.1", port=9999)` in `server.py` | | No data appears | Verify query and Snowflake table permissions | ## Offline Mode If Snowflake is temporarily unavailable: 1. Click **↓ import .json** in the dashboard 2. Select a previously exported `.json` file 3. The dashboard will show local data while Snowflake is down ## Next Steps - **Monitoring**: Set up health checks for Snowflake availability - **Caching**: Add Redis caching layer for high-traffic periods - **Auth**: Integrate Okta or Entra ID for production deployment - **Deployment**: Deploy to Azure Container Apps or Netlify Functions ## Related Resources - [Snowflake Python Connector Docs](https://docs.snowflake.com/en/developer-guide/python-connector) - [Flask Documentation](https://flask.palletsprojects.com/) - [D3 Gallery](https://observablehq.com/@d3/gallery)