Skip to main content

Allye Base

Allye Base is the core foundation of the Allye platform, equipping you with essential tools for data acquisition, transformation, interactive & connected visualization, and multivariate statistical analysis.

Key Features

1. Data Management

Connect to various data sources, from local files to cloud databases, and manage your datasets seamlessly.

WidgetDescription
FileImport data from diverse file formats including CSV, Excel (xlsx), TSV, and Tab-delimited files.
DB ConnectorConnect directly to databases. Supports BigQuery, Redshift, SQL Server, PostgreSQL, and MySQL.
SQL ExecutorExecute custom SQL queries directly against your connected databases.
Google SheetImport data directly from Google Spreadsheets (requires authentication).
Allye Data Receiver / TransmitterExchange data seamlessly between Allye and external Python programs.
Save DataExport and save your processed data as CSV or Excel files.
Data TableView and inspect the raw data in a spreadsheet-like format.

2. Visualization

Explore your data with a wide range of interactive charts. Allye's visualization widgets are "connected," meaning selections in one chart automatically propagate to others.

CategoryWidgets & Description
Basic VisualizationBar Plot, Line Plot, Box Plot, Scatter Plot, and Distributions (Histogram).
Advanced Visualization3D Scatter Plot: Visualize data in three-dimensional space.
Feature Statistics: Inspect basic statistical properties of data features.
Visualize Connector: Acts as a hub to synchronize selection states across multiple visualization widgets. See Connected Visualization for details.

3. Data Transformation & Preprocessing

Clean, shape, and enrich your data before analysis using powerful ETL tools.

WidgetDescription
Python NotebookExecute arbitrary Python code within an integrated Jupyter Notebook environment.
Select Rows / ColumnsFilter data by selecting specific rows or columns based on conditions.
Group By / Pivot TableAggregate data, calculate summaries, and reshape datasets using pivot logic.
MergeJoin multiple tables using standard SQL-style joins: Inner, Left, and Outer joins.
FormulaCreate new variables or modify existing ones using mathematical formulas.
PreprocessPerform extensive data cleaning, including normalization, missing value imputation (removal/interpolation), and discretization/continuization.
Edit DomainRename columns, change data types, and modify variable attributes.
Data SamplerRandomly draw a subset of data points for training or testing.

4. Statistical Analysis

Generate comprehensive statistical reports and build models with detailed diagnostics.

Analysis TypeDescription & Outputs
AB TestCompares groups across multiple target metrics, and reports lift, confidence intervals, and p-values. Automatically selects an appropriate test (T-test / Chi-square / Mann–Whitney U) based on the target variable type, with optional multiple-comparison correction (Bonferroni / Holm / Benjamini–Hochberg).
Regression AnalysisInteractive linear regression (with optional L1/L2 regularization) and a full statistical report (not just predictions).
Metrics: R², Adjusted R², MSE, RMSE, AIC, BIC.
Model insight: Coefficients, confidence intervals and p-values (available when unregularized), and optional VIF to check multicollinearity.
Diagnostics: Predicted vs. Actual and residual plots (Residuals vs Fitted, Normal Q-Q, Scale-Location, Residuals vs Leverage).
Binary AnalysisLogistic regression for binary outcomes (with optional L1/L2 regularization), designed for both interpretability and evaluation.
Metrics: Accuracy, Precision, Recall, F1, AUC, Log Loss.
Interpretation: Coefficients, odds ratios, and (when unregularized) confidence intervals and p-values; optional VIF for multicollinearity checks.
Plots: ROC curve and confusion matrix.
PCAPrincipal Component Analysis for in-depth understanding and dimensionality reduction.
Selection: Choose a fixed number of components or target explained variance; optional normalization.
Outputs: Transformed data, component loadings, and explained-variance summaries.
Visualizations: Scree plot, loadings table/plot, and biplot.

Templates

To see practical examples of analysis using Allye Base, please refer to the EDA Tutorial section.