Build your own Amazon analytics app, on Nova data.

Build an Amazon seller analytics app connected to Nova Data Share.

## Step 1 — Ask me these questions before doing anything else

Ask me the following in one message, then wait for my answers:

1. What is your company website URL?
   (I will copy your brand colors, fonts, and logo from it)

2. What should the app be called?

## Step 2 — Secrets setup

Before writing any code, prompt me to fill in the following secrets via the "Update secrets" panel:

- CLICKHOUSE_HOST
- CLICKHOUSE_USER
- CLICKHOUSE_PASSWORD
- CLICKHOUSE_DATABASE

These will be used server-side only. Never expose them in the frontend.

Important: warn the user that CLICKHOUSE_USER must be kept exactly as provided, including the mysql4qjne6q9i1y_ prefix. The connection will fail without it.

## Step 3 — Design

Scrape the website URL I provided and extract the visual identity: colors, fonts, logo, and overall aesthetic. Apply it consistently across the entire app so it feels like my own branded product.

## Step 4 — Starter dashboard

Build a first dashboard as a starting point with:

### KPI cards (top row, last 30 days)
- Total revenue (product_sales)
- Net profit (net_profit)
- Units sold (units)
- TACOS (ppc_spend / product_sales * 100)
- Return rate (returned_units / units * 100)

### Charts
- Daily trend: product_sales + net_profit + ppc_spend (line chart)
- Revenue by marketplace: grouped bar chart (marketplace_name)
- Top 20 ASINs by net_profit: sortable table with columns: child_asin, units, product_sales, ppc_spend, ACOS, margin %

### Filters
- Date range picker (default: last 30 days)
- Marketplace selector (multi-select)
- FBA / FBM / All toggle (is_fbm)

## Database schema reference — consult this before building any feature

Connection:
- Protocol: MySQL wire, port 3306, SSL required
- One table: nova_share.product_metrics
- Grain: one row per seller x marketplace x child ASIN x SKU x day
- Data refreshes every hour from Nova's pipeline
- Always filter on metric_date. Never run unfiltered queries
- Always GROUP BY