Visualization
DataMosaicA web-based data visualization and dashboarding platform: connect to sample datasets, uploaded spreadsheets, or live REST APIs, build drag-and-drop dashboards out of bar/line/pie/scatter/heatmap/map/table/KPI widgets, and collaborate on them in real time — comments, live presence, and instant sync of every edit to everyone else looking at the same dashboard.
Every dashboard widget ultimately plots one of five source types, all under Data Sources:
From Home, click + New dashboard for a blank canvas, or open an existing one. Dashboards are a 12-column, drag-and-resize grid — click + Add widget to place one.
For a SQL Database source, the widget editor shows a node-and-edge Query graph instead of a column picker. Add Table, Join, Filter, Aggregate, Sort, and Limit nodes, then click a node's right-hand dot and another node's left-hand dot to wire them together (click a wire to remove it). Run preview shows the result in the editor before you save.
The whole graph compiles to one read-only, parameterised WITH … SELECT that runs in the database — only the result rows come back to DataMosaic, which shapes them for the chart. Open Generated SQL under the graph to see exactly what will run. Filter conditions support the usual comparison operators plus not in (compiled to a bound NOT IN (…) list).
Dashboard cross-filters work here too: clicking a category on one widget adds a bound WHERE to the other SQL widgets' final select, so the filter applies before the 10,000-row result cap rather than trimming rows afterwards.
The Add widget dialog picks a chart type, a data source, and how to map its columns — category/value axes, an optional split-by series, and an aggregation (sum, avg, count, ...). Bar and line widgets can also show a linear-regression trend line or flag statistical outliers (more than two standard deviations from the mean).
Chart types are grouped into five tabs. Bar, Line, Area and Scale hold the Chart.js variants — for example vertical, horizontal, rounded, floating, stacked and stacked-group bars, plus time-scale charts that sort and space points by date. Other holds Map (Leaflet + OpenStreetMap), Heatmap, Table, KPI, and the remaining types. Each variant shows only the field pickers and options it needs (a second measure, a stack group, a bubble-size field, and so on), and the original eight type keys (bar, line, pie, scatter, heatmap, map, table, KPI) keep their meaning, so existing dashboards render unchanged.
A Table widget has a Columns picker in the editor so you choose which columns appear. If the source or query behind a table changes and some saved columns no longer exist, the table shows only those that still do — or every column if none match — rather than rendering blank cells.
The Colors button in the dashboard toolbar sets a theme for every chart on that dashboard: Default, Ocean, Sunset, Forest, Grape, Slate, High contrast, Pastel, or Custom… to pick your own colors. A widget can override the dashboard theme from its own editor. Themes also color trend lines and anomaly highlights, and they carry through to public share links and embeds. A dashboard with no theme renders exactly as before.
A finished dashboard mixes several widgets freely — clicking a bar or pie slice cross-filters every other widget on the same dashboard by that category.
Click + New DBOps Dashboard on Home and pick a Shared Data Connection to get a one-click "fleet health" view: DataMosaic provisions a dashboard plus two DBOps Telemetry sources (connection utilization and unused indexes) plus a line widget and a table widget, all pre-wired, in one call.
Every open dashboard is a Socket.IO room: comments, collaborator presence, and every widget/layout edit sync live to everyone else looking at it, with no manual refresh. The right-hand panel switches between Comments, Collaborators, Share, Snapshots, and Git.
The Collaborators tab has a name picker listing people who have signed in to DataMosaic (minus the owner, you, and anyone already added). Pick a person and a role — viewer or editor — and change or remove it later from a per-collaborator dropdown. The server validates every role, rejects the owner or unknown users, and audit-logs sharing, role changes and unsharing. Dashboards shared with you carry a Shared with you badge on Home.
The Share tab (owner or admin) turns on a public link and manages it:
All share changes are written to the audit log.
Unlike Detanami's or Smith's chat panels, DataMosaic has no single query or command target — so its AI Assistant (top right, on every page) is a general Q&A assistant aware of your data sources and dashboards by name and type, with no insert/apply action. It's the one AI DBA Agent surface that's pure conversation.
Per-widget CSV export works for any widget; chart widgets also export PNG directly. For the whole dashboard, use the Export menu in the toolbar:
| Option | What you get |
|---|---|
| Image (PNG) | The whole dashboard as one image. |
| The whole dashboard as a PDF. Image and PDF are both rendered in your browser (the libraries load on demand). | |
| Data workbook (.xlsx) | One sheet per widget, honouring any cross-filter currently applied. |
| Dashboard file (.json) | A portable definition — layout, widgets, theme — to re-import elsewhere or keep as a backup. |
| Print… | The browser's print dialog against a dedicated print stylesheet. |
Import on the Home page takes a dashboard file and creates a new dashboard from it, re-linking data sources by name and type. Dashboards can also be exported straight from their card on Home. Exports and imports are audit-logged.
DBMojo Docs