Allye Documentation
Why Allye? - Hybrid Analysis Environment
Built upon the foundation of open-source Orange Data Mining, Allye is a no-code/code hybrid platform engineered for professional rigor. It combines the intuitive nature of visual programming with the flexibility of raw code, all while running locally, securely, and quickly (in-memory processing).
Breaking the Compromise
Traditionally, analysts had to choose between the ease of GUI tools and the flexibility of coding. Allye eliminates this trade-off, offering a unified environment that serves both engineers and business analysts.
| Traditional Tools | (Vibe) Coding | Allye | |
|---|---|---|---|
| Pros | Speed, Easy | Flexibility | Hybrid |
| Cons | Limited Flexibility | Limited Speed, Complex |
Core Capabilities
1. Hybrid Environment
- Flexibility meets Efficiency: Start your analysis with code (like Jupyter) and instantly drop it into a visual workflow or vice-versa. You can seamlessly switch between writing logic and using GUI components.
2. Interactive & Connected
- Inter-connected Views: Allye Visualizations are all connected and interactive. Selecting data points in a scatter plot immediately highlights the same data across tables and histograms, revealing hidden patterns instantly.
- Structured Workflows: Your trial-and-error process is preserved as a visual flow, making it easy to retrace steps or explain your logic to stakeholders.
3. Automated Rigor: Quality Control Built-in
- Automatic Diagnostics: Necessary validation metrics appear automatically—such as scree plots, biplot, loading barplot for PCA or balance checks, love plot for Propensity Score Matching, p-value, confidence interval, VIF and residual analysis for Regression Analysis—so you never miss critical sanity checks.
- Reliable Reporting: The system highlights abnormal values (for example, when a p-value exceeds 0.05) to guide you toward correct analysis.
4. Causal Inference: From "What" to "Why"
Business decisions require understanding cause and effect, not just correlations. Allye makes Causal Inference a core capability.
- Beyond A/B Testing: We provide a comprehensive suite of quasi-experimental methods (PSM, DID, DML, Causal Forest) for scenarios where randomized control trials are impossible.
- Structured Discovery: Explore causal structures and verify hypotheses visually without needing complex script implementation.
Security & Performance
Local. Secure. In-Memory.
Allye is designed for environments where data privacy is paramount.
- Local Execution: All processing happens on your local machine or on-premise server. Your data never leaves your environment.
- In-Memory Speed: High-speed in-memory processing ensures that your train-of-thought is never interrupted by loading times.
Powered by Open Source: Allye is built upon the robust foundation of Orange Data Mining (and other various oss), enhanced with our proprietary AI technologies.