From CSV Upload to Interactive Dashboard in Under 2 Minutes

You have a CSV file with sales data. You need a quick dashboard for your team meeting in 10 minutes. No time for Jupyter notebooks, no time for Tableau. Here is how to go from raw CSV to interactive dashboard in under 2 minutes with Harbinger Explorer.

Prerequisites

• A Harbinger Explorer account (free trial works)
• A CSV file (we will use a sample sales dataset)

Step 1: Upload Your CSV

After logging in, you land on the dashboard. The hero section has a drag-and-drop zone front and center.

1. Enter a table name in the input field (e.g., sales_data)
2. Drag your CSV file onto the drop zone — or click to browse

The system parses your file instantly and shows a column mapping preview:

Review the detected types, adjust if needed, then click Confirm & Load.

Step 2: Check for Sensitive Data

Harbinger Explorer automatically scans for PII (Personally Identifiable Information). If your dataset contains columns that look like emails, phone numbers, or social security numbers, you will see a warning:

"Column 'customer_email' detected as PII (email pattern). Pseudonymize before sharing with AI?"

Click Pseudonymize to replace real values with consistent fake ones. The AI agent will work with the masked data, but your original values remain intact in DuckDB for your own queries.

If your sales dataset has no PII, you will not see this step — the system only flags what it detects.

Step 3: Ask the Agent for Insights

Navigate to the Agent Chat (sidebar or the "Ask the Agent" button). Your uploaded table is automatically available. Try these prompts:

"Show me monthly revenue trends for 2026"

The agent generates SQL, executes it against your DuckDB table, and produces a line chart showing revenue by month.

"Which region has the highest average order value?"

The agent runs a GROUP BY query and creates a bar chart comparing regions.

"Give me a breakdown of revenue by product category"

You get a donut chart with percentages and absolute values.

Step 4: Pin Charts to Your Dashboard

Each chart the agent generates has a Pin button in the top-right corner. Click it to save the visualization to your personal dashboard.

After pinning 3-4 charts, go to the Insights page to see your dashboard:

• Monthly revenue trend (line chart)
• Revenue by region (bar chart)
• Product mix (donut chart)
• Top 10 products by quantity (horizontal bar chart)

The dashboard updates automatically when you refresh the underlying data. Upload a new CSV with the same table name, and your charts reflect the new numbers.

Step 5: Export and Share

Need to share the results? Harbinger Explorer offers several export options:

• CSV — Download the query results as a CSV file
• SQL — Export the generated SQL queries for your pipeline
• Parquet — Columnar format for downstream processing

Click the Export button on any query result or chart to choose your format.

Bonus: Write Your Own SQL

If the agent's suggestions are not enough, switch to the Workspace tab. The Monaco SQL editor gives you full DuckDB SQL with auto-complete:

SELECT
  DATE_TRUNC('week', date) AS week,
  region,
  SUM(revenue) AS weekly_revenue,
  SUM(revenue) / SUM(quantity) AS avg_price
FROM sales_data
WHERE date >= '2026-01-01'
GROUP BY ALL
ORDER BY week, region;

Results appear instantly — DuckDB WASM processes your data right in the browser.

What We Covered

1. Upload — Drag-drop CSV with automatic schema detection
2. PII Check — Automatic detection and pseudonymization of sensitive data
3. Agent Insights — Natural language queries that produce charts
4. Dashboard — Pin charts for a reusable dashboard
5. Export — CSV, SQL, or Parquet for sharing

Total time: under 2 minutes. No server setup. No configuration files. Just data and insights.

Try it yourself at harbingerexplorer.com — the free trial gives you full access to all features.