Categories
Self-Hosted Data Visualization Business Intelligence Dashboards Data Analysis Exploratory Data Analysis Causal Analysis Automated InsightsStars
Forks
Watchers
Developer links
Rath
RATH, an automated data exploration workbench, turns any table into a gallery of charts worth reading, without building each chart by hand. Start with a CSV, JSON, or Excel file, or pull tables from MySQL, PostgreSQL, Oracle, SQL Server, SQLite, or Airtable. Eight built-in demo datasets are ready to try. Next comes a profile: every field gets a distribution chart and summary statistics. A metadata view retags fields as dimensions or measures. Data Autopilot then searches the dataset and pages through generated insight charts. Charts can be starred, exported, or opened in the editor. Starting from a single field, Data Copilot lists associated patterns to follow, with controls to add variables, filters, and computations. For hands-on work, the drag-and-drop chart builder offers fifteen chart types, including heat maps, box plots, and maps, plus a button that applies the recommended chart. With Data Painter, you brush points into up to ten labeled color groups or erase anomalies, then search for a chart that explains the painting. Causal analysis runs one of ten discovery algorithms, including PC, FCI, GES, and GRaSP. It then draws a directed graph with influence percentages, plus charts and statistics for any field you select. Dashboards combine titles, text, and charts in movable, resizable cards with four themes and global filters. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. AGPL-3.0 licensed.
Benefits
- Automatic Insights From Any Table
- Autopilot scans the loaded table and pages through generated charts, while Copilot lists associated patterns for a chosen field, so exploration starts from findings instead of a blank canvas.
- Connect Files and Databases
- Load CSV, JSON, or Excel files, pull Airtable tables, or query MySQL, PostgreSQL, Oracle, SQL Server, and SQLite through the bundled connector service running on your own server.
- Manual Charts for Full Control
- A drag-and-drop editor offers fifteen chart types, filters, and color, size, and shape channels, so any automatic suggestion can be refined into the exact view you want to share.
- Self-Hosted With Docker Compose
- Docker Compose builds the web client, serves it through nginx on port 9083, and starts the connector service beside it, so one server hosts the whole workbench.
Features
- Data Autopilot Insights
- Generates pages of insight charts from the loaded table, with a view search box and edit-chart, data-painter, associate-views, star, and export actions on each result.
- Data Copilot Suggestions
- Starts from one field and lists associated patterns, each with related-insights, star, and edit-chart actions, plus controls to add variables, filters, and computations.
- Drag-and-Drop Chart Editor
- A field list, filters, and color, opacity, size, shape, and detail channels drive fifteen chart types, and an auto-viz panel applies the recommended chart.
- Data Painter
- Brush chart points into up to ten labeled color groups, erase anomalies, adjust brush size and sample percentage, then search for a chart explaining the painting.
- Causal Discovery Graphs
- Runs one of ten discovery algorithms, draws directed edges, and offers graph and matrix views with per-field charts and statistics for any selection.
- Dashboards With Themes
- Combine titles, text, and charts in movable, resizable cards, with transparent, outline, dropping, and neumorphism themes plus global filters.