Auto-Generated Dashboards from Data Sources

3 min read Updated May 21, 2026

Every connected source gets an auto-generated insight dashboard via clariBI's preprocessing pipeline. Here's what gets analyzed, where to find the dashboard, and how to customize it.

# Auto-Generated Dashboards from Data Sources Every time you connect a new data source or upload a file, clariBI's preprocessing engine analyzes the data and creates an insights dashboard automatically. No manual chart configuration required. ## How It Works ### The Preprocessing Pipeline 1. **Data ingestion** — You upload a CSV, connect a database, or authorize an integration (Google Ads, Jira, Meta Ads, Stripe via MCP, HubSpot via MCP, etc.). 2. **Schema detection** — clariBI inspects columns, data types, and value distributions. 3. **Statistical analysis** — The engine calculates summary statistics, identifies time-series columns, detects categorical distributions, and spots outliers. 4. **Chart generation** — Based on the analysis, clariBI selects the most appropriate chart types and builds pre-computed visualizations. 5. **Dashboard creation** — A new dashboard named **"[Source Name] - Data Insights"** appears in your dashboard list with all generated charts. ### What Gets Analyzed - **Numeric columns**: Aggregations (sum, average, min, max), distributions, and trends over time - **Categorical columns**: Value counts, top categories, and proportional breakdowns - **Date/time columns**: Time-series trends, period-over-period comparisons - **Relationships**: Correlations between numeric fields, category-metric breakdowns ## Finding Your Auto-Generated Dashboard 1. Go to **AI Dashboards** in the sidebar 2. Look for a dashboard named **"[Your Source Name] - Data Insights"** 3. Auto-generated dashboards are created one per data source If you rename the data source later, the dashboard name updates automatically to stay in sync. ## Available Chart Types Auto-generated dashboards draw from 24 visualization types including: - Bar charts and line charts for trends and comparisons - Pie charts and treemaps for proportional data - Scatter plots for correlations - Heatmaps for multi-dimensional patterns - Tables for detailed breakdowns - Gauge and KPI cards for key metrics The engine picks the chart type that best fits each column or column pair in your data. ## Customizing Auto-Generated Dashboards Auto-generated dashboards are fully editable: - **Rearrange widgets** — Drag and drop to reorder - **Remove charts** — Delete any widget you do not need - **Add new widgets** — Add your own custom charts on top of the generated ones - **Change chart types** — Switch a bar chart to a line chart if you prefer - **Adjust filters** — Add date ranges or category filters Changes you make are saved and preserved when the data refreshes. ## Refresh Behavior When your data source syncs new data (on its configured schedule or via manual refresh), the preprocessing pipeline runs again: - Existing auto-generated charts update with fresh data - New columns or changed data patterns may produce updated visualizations - Your manual customizations (rearrangements, added widgets) are preserved ## Deleting a Data Source If you delete a data source, the auto-generated dashboard is **not deleted**. It remains viewable with its last known data, and a banner indicates the source has been disconnected. This protects your work — you can always reconnect a replacement source later. ## Availability by tier - **Free**: auto-generated data insight dashboards are accessible on every uploaded file or connected source. - **Trial / Starter / Professional / Enterprise**: all features above plus AI-powered conversational analytics on top. - **Lite**: auto-generated dashboards via the manual deterministic engine (no LLM calls). ## Tips - Upload a clean CSV with headers and consistent data types for the best automatic results - Connect multiple sources and compare their auto-generated dashboards side by side - Use the auto-generated dashboard as a starting point, then refine it for your specific reporting needs - Ask clariBI questions about your data via **Conversational Analytics** for even deeper analysis beyond what the auto-dashboard surfaces