Wizard - Statistics, Visualization, Data Analysis, Predictive Modeling
Wizard is a new Mac app that makes data analysis easier than ever. No programming, no typing — just click and explore. Wizard makes statistics accessible to beginners, but beneath the surface lies a full set of tools for doing professional research.
Featured by Apple! "New and Noteworthy" (Feb. 2013) and "Apps for Business" (Apr. 2013)
"You can be looking at statistical analyses less than a minute after importing your data." —MacFormat Magazine (UK)
It's true: within minutes of importing your first data set into Wizard, you'll be generating graphics, uncovering correlations, producing p-values, building models, and sharing insights with colleagues. Most customers find that the program pays for itself in just a few days of use. Explore Wizard's delightful feature set, including...
Wizard will start graphing your data as soon as you click on it. It uses native Mac graphics to give you crisp visual summaries — box-plots, scatterplots, histograms, survival curves, and more — in a fraction of a second.
EASY STATISTICAL MODELING
Build sophisticated statistical models in seconds with Wizard's intuitive modeling interface. Regression estimates are instantly recomputed as you play around with the controls and experiment with new specifications.
Wizard supports drag-and-drop, multi-level Undo, and other interface conveniences that Mac users expect from their software. An interactive tutorial will help you get started.
Dozens of functions are available to help you get data into the form you want. Compute quantiles, perform day-of-week calculations, multiply columns together, and more.
Export colorful graphics as web-friendly PNG or print-quality PDF. If you're surrounded by PC users, export your models as interactive spreadsheets that can be used to run "What if?" scenarios in any version of Microsoft Excel.
LIST OF STATISTICAL TESTS
Wizard supports the most common statistical tests and models, including...
+ Shapiro-Wilk test of normality
+ 1-sample Kolmogorov-Smirnov (normality and uniformity)
+ Pearson's goodness-of-fit (equal proportions)
+ Pearson's goodness-of-fit (chi-square)
+ t-test (paired and unpaired)
+ ANOVA (1-way, 2-way, and repeated measures)
+ Correlation and R²
+ Median tests: Mann-Whitney and Kruskal-Wallis
+ NEW: Wilcoxon signed-rank and Friedman tests
+ Survival analysis: Log-rank test
+ 2-sample and N-sample Kolmogorov-Smirnov
+ Linear regression (OLS)
+ Weighted linear regression (WLS)
+ Poisson and geometric regression
+ Logistic regression (Logit) and Probit
+ Multinomial Logit and Ordered Probit
+ Negative Binomial (NegBin-2)
+ Cox Proportional Hazards
+ Fixed effects
+ Robust standard errors
+ Clustered standard errors
+ Joint significance tests (Wald tests)
+ Odds ratios
+ Residual analysis
+ Sensitivity/specificity analysis (ROC curves)
+ Censoring of survival data (Cox models)
+ Interactive prediction assistant
+ Copy models as R code
Is there a missing feature you'd like to see? Just click "Email the Developer" from the Help menu and send me your thoughts.
LIST OF FILE FORMATS AND DATABASES
Although you have the option to enter data manually, more likely you'll want to import data from one of the following sources:
• Excel (.xls/.xlsx)
• Numbers ('09 and later)
• R workspace and single-object files (.RData and .rds)
• MS-Access (.mdb/.accdb)
• Plain text (Comma-, tab-, and custom-delimited values)
To import and export other formats, including SPSS, Stata, and SAS, you can either buy In-App Purchases inside Wizard (Wizard > Unlock Pro Features...), or purchase Wizard Pro, also available in the Mac App Store. The set of included importers is the only difference between the Pro and Standard versions.
• Fix several crashes
◦ Viewing confidence intervals of the mean on 10.7 and 10.8 caused a crash
◦ Creating a custom covariate partition in the Summary view caused a crash
◦ Very long text values sometimes caused a crash
◦ Proportional Hazards models with a specified cause of failure caused a crash
• New features for Contingency Table Analysis
◦ McNemar Test (2 repeated binary measures)
◦ Cochran's Q (3 or more repeated binary measures)
◦ Stuart-Maxwell Test (2 repeated category measures)
• "Show Estimation Command" now works in the Summary view, too. (Generates commands for reproducing results in R)
• Pivot tables have improved support for hierarchical data
• Column > Subtract now works on date columns
• Summary view supports cross-column comparisons of category and survival-time data
• Proportional Hazards models are much faster
[Older release notes below]
New in 1.5.0:
Our mad scientists have been feverishly slaving away on the latest version of Wizard in order to make the art of data analysis easier than ever. Enjoy all the great new features in Wizard 1.5: (Scroll down to see what's new in 1.5.5)
• A new "Treat as Survival Times" option, which lets you produce survival (Nelson-Aalen) curves, view confidence bands, and perform log-rank tests. Survival analysis is perfect for analyzing failure, relapse, and mortality data, as well as other kinds of event data.
• Paired t-tests and repeated measures ANOVA: To analyze "before and after" data, just selecting multiple columns in the Summary view and checking the box labeled "Repeated Measures"
• Text partitions: Now you can group non-category text data into groups, e.g. A-G, H-M, and N-Z.
• Export image data: The data behind all the graphics in the Summary view can now be exported as CSV or XLS.
• Zoom to filtered values: There's a new magnifying glass button in the top-right corner of the Data Filters pane that lets you zoom all the graphics to the filtered values.
• Improved prediction intervals: The Predict view will now draw prediction intervals in the response curve, in addition to displaying the numeric values.
• Interaction terms: Models can now include separate interaction terms between two explanatory variables. For continuous variables, this is the product of the two terms; for category variables, this is an extra set of dummy variables representing all possible combinations of the two terms. Interaction terms will "just work" in the Predict view and in export XLS models.
• Reorder outcome categories: If your outcome categories aren't ordered properly, now you can just drag and drop to rearrange them. This should make Ordered Probit models more immediately useful without having to create ad-hoc columns.
New in 1.5.1:
• Keep popovers open as you switch windows.
• New option to hide charts in the Summary view (View > Hide Summary Charts)
• Fix a crash when using a date/time column as a covariate
• Fix a crash when importing values of infinity or negative infinity
New in 1.5.2:
• Improved support for missing values in RData files
• CSV importer ignores leading blank lines
• RData files can have extension .rda
• Fix oddities with ROC curves when frequency weights are applied
New in 1.5.3:
• Support for Numbers 3.5 documents
• New repeated-measures tests: Wilcoxon signed-rank test for 2 numeric columns, Friedman test for 3 or more numeric columns
• Many speed improvements throughout the program
• Fix a freeze when customizing filters
• Some date/time values in CSV files were ignored
• Preserve date/time semantics when creating a new table from a Pivot table
• Fixed-effect models exported as XLS didn't work properly
New in 1.5.4:
• Survival curves: dotted lines for censored values, and more grid lines for large graphs
• Accept percentage values in CSV files
• Support for importing DBF (dBase) files
Share with Others
- Last changed:
- Nov 11, 2014
- Evan Miller
- 5.1 MB
- Current Version Average Rating:
- 5.00 (1)
- All Versions Average Rating:
- 5.00 (51)