MIDAS Documentation

MIDAS is a free exploratory data analysis tool that runs entirely in your browser. From basic analysis — seeing the distribution of your data and the relationships between variables with histograms, scatter plots, and summary statistics such as means and quantiles — to statistical modeling, you can do it all through point-and-click in the UI. Modeling includes linear regression, generalized linear models (GLM), generalized linear mixed models (GLMM), ANOVA, principal component analysis (PCA), random forests, and survival analysis.

Everything you load is processed inside your browser and is never sent to an external server.

Getting Started

• Basic Usage - Walk through the basic operations, from loading data to saving results, with sample data

User Guide

Data Preparation

• Data Preparation and Import - How to load CSV/TSV files
• Datasets - Managing imported data and derived datasets created through SQL or other transformations
• Sample Datasets - Description of sample data included in MIDAS

Data Exploration

• The Data Table Tab - Viewing, filtering, and sorting data
• How Row Selection Works - How row selection works across tabs
• The Selected Rows Tab - Viewing selected rows and saving them as a derived dataset
• The Filtered Data Tab - Viewing and refining rows behind double-clicked graph or crosstab elements

Screen Layout

• Workspace and Layout Management - Managing multiple analysis tasks in parallel

Data Processing

• The Convert Column Types Tab - Converting data types and handling errors
• The Reshape Tab - Converting between Wide and Long formats
• The Dummy Coding Tab - Converting categorical variables to dummy variables
• The SQL Query Editor Tab - Transforming data using SQL
• The Manage Enums Tab - Creating and managing Enum types for categorical data
• The Orthogonal Polynomials Tab - Generating orthogonal polynomial columns from a numeric column

Data Visualization

• The Graph Builder Tab - Histograms, scatter plots, bar charts, time series plots, pair plots, and more
• Advanced Graph Creation - Layering multiple graph types and faceting based on Grammar of Graphics

Statistical Analysis

• The Statistics Tab - Mean, standard deviation, quantiles, and other summary statistics
• The ANOVA Tab - One-way and two-way analysis of variance with Tukey HSD post-hoc comparisons
• The DoE Analysis Tab - 2-level orthogonal array generation and factorial experiment analysis
• The Crosstab Tab - Pivot tables for categorical variables
• The Linear Regression Tab - Ordinary least squares regression with coefficient confidence intervals and residual diagnostics
• The GLM Tab - Logistic, Poisson, and other regression models for response variables that need not follow a normal distribution
• The GLMM Tab - Random intercept models for grouped data
• The Kaplan-Meier Tab - Estimating survival curves and comparing groups with RMST
• The Cox Regression Tab - Estimating covariate effects with Cox proportional hazards models
• The Random Forest Tab - Random forests for classification and regression with feature importances
• The PCA Tab - Dimensionality reduction and exploring correlation structures among variables

Organizing Analysis Results

• Reports - Saving graphs and statistical results together
• Export - Export data as CSV/TSV/JSON, download graphs as SVG

Project Management

• Project Management - Managing datasets, reports, and models
• The Project Overview Tab - View and manage resources
• The Project Lineage Tab - Visualize dependencies
• The Compare Project Tab - Compare project versions
• MDS Files - Saving, exporting, and signing project files
• Storage Management - Viewing and deleting saved projects and checking storage usage
• Managing Signing Keys - Verifying MDS file signatures and managing trusted keys

Reference

• Custom Graph Reference - Geometry/Statistics list
• Agent API (window.midas) - Controlling MIDAS from AI agents and external tools
• Numerical Accuracy - Verifying statistical computation accuracy with NIST Statistical Reference Datasets

Tutorials

Step-by-step walkthroughs with sample data, each starting from a question you want the data to answer and proceeding through method choice to interpreting the results.

• Comparing Groups with ANOVA and Linear Regression - A complete workflow that separates group differences from environmental factors, using production line inspection data
• Optimizing Injection Molding Conditions with Design of Experiments - Analyze a 3-factor 2-level factorial experiment
• Survival Analysis with the Kaplan-Meier Method - Estimate survival curves and compare groups
• Grouped Binomial GLM with Dose-Response Data - Logistic regression for aggregated binomial data

Statistical Concepts

Background knowledge on the statistical methods used in MIDAS. Refer to these pages as needed.

• Data Types and Measurement Scales - Nominal, ordinal, interval, ratio scales and their impact on analysis
• OLS Fundamentals - Normal equations, Gauss-Markov theorem, VIF
• GLM Fundamentals - Exponential family, link functions, IRLS
• Survival Analysis Fundamentals - Censoring, Kaplan-Meier, Cox proportional hazards model
• GLMM Fundamentals - Random effect models, REML, BLUP, ICC
• Numerical Computing Fundamentals - How floating-point precision, catastrophic cancellation, and condition numbers affect computation accuracy
• Missing Data Mechanisms - MCAR, MAR, MNAR, and the assumptions behind listwise deletion
• Glossary - Definitions of estimator, convergence, likelihood, deviance, and more

System Requirements

• Privacy and Security - Data processing, storage, external communication, and browser requirements
• PWA and Offline Use - Install as an app and work offline

Support

• Release Notes - New features and change history
• For questions or bug reports, contact contact@midas-app.org

Also available as a Markdown file.