Geographic work, data mining and extraction, and other specialty statistics software for the Mac

These are other Macintosh software packages used for data analysis beyond those for regression, ANOVA, or time series analysis.

Updated 10/9/2021

Data extraction software estimates data based on image files such as scanned charts and graphs. Carl Witthoft suggested this category and provided us with several starting programs. Also see graphing / data visualization software; and general statistics packages. Page fully updated January 2021.

Data extraction

Engauge Digitizer

Current version: 12.1
Free. Signed. Catalina-ready.
Via GitHub.
Listing updated: 10/3/2021; software updated in 2019

"...converts an image file showing a graph or map, into numbers. The numbers can be read on the screen, and written or copied to a spreadsheet." Can remove gridlines, match points, trace curves, match axes, handle a wide variety of graph types, and has other nifty features. Linux, OS X, and Windows versions; it’s on the Mac App Store.

Engauge Digitizer takes images in PNG, JPG, and TIF format and recovers the data points from graphs, possibly to be used in new graphs. Work can be saved in DIG format for later editing. Engauge Digitizer has numerous special features to make data more accurate and easier to obtain; it’s pretty impressive.

DataThief III

Current version: DataThief III 1.7; Java program
$25 (version 1 was postcard/beerware)
Listing updated 10/3/2021. Software updated 2015.
Unsigned. Runs under Catalina.

DataThief reverse engineers data from a scanned plot, so you can incorporate published data in your plots—very handy if you need to compare your data with that of an article that doesn’t provide it in a table.

The current crossplatform version, using Java, can trace most continuous lines, even when they cross themselves, and can convert numbers to other formats (e.g. dates). Version III works on MacOS 8 and 9 as well as X and those other platforms. DataThief II (version 1.21), for older Macs, is still available on the DataThief web site.

The original version of DataThief was written by Kees Huyser and Jan van der Laan. Available from Bas Tummers.

Geographic resources

GRASS

Configurations Available: Mac, Linux, Windows
Price: Free (and open source)
Current Version: 7.8.5 / 8.0 Preview
Listing updated 10-3-2021
Last software update, 12-21-2020

Michael Barton pointed out that GRASS is used for geographic resources data management, image processing, graphics production, spatial modelling, and visualization of many types of data. It is an official project of the Open Source Geospatial Foundation.

Originally developed by the Army as a tool for land management and environmental planning, GRASS is a powerful utility with a wide range of applications in many different areas of scientific research. GRASS is currently used in academic, government, and commercial settings. Attributes are managed in a SQL-based DBMS.

GRASS 6 added a new topological 2D/3D vector engine and support for vector network analysis. A new display manager has been implemented. The NVIZ visualization tool was enhanced to display 3D vector data and voxel volumes. Messages are partially translated with support for FreeType fonts, including multibyte Asian characters. New LOCATIONs can be auto-generated by EPSG code number. GRASS is integrated with GDAL/OGR libraries to support an extensive range of raster and vector formats, including OGC-conformal Simple Features.

Quantum GIS

Configurations Available: Mac, Linux, Windows
Current version: 3.20
Price: Free (and open source)
Signed for newer Mac versions
Listing updated 10-3-2021; last release, 9-10-2021

Quantum GIS is a somewhat less powerful but easy to use GIS package for Mac, Linux, and Windows. It is also an Open Source Geospatial Foundation project, and it supports numerous vector, raster, and database formats and functions.

Generic Mapping Tools (GMT) for geographic data

Configurations Available: Mac (under X11), Linux, UNIX, OS/2, Windows [requires Cygwin or VirtualBox for full function]
Current Version: 6.2.1 (released June 2021)
Price: Free (and open source)
Listing updated 10-3-2021

Generic Mapping Tools, or GMT, is an open source collection of many tools for manipulating geographic and Cartesian data sets (including filtering, trend fitting, gridding, projecting, etc.) and producing Encapsulated PostScript File (EPS) illustrations ranging from simple x-y plots via contour maps to artificially illuminated surfaces and 3-D perspective views. A MATLAB extension is available.

Time Series Software

gretl

gretl does time series and other statistics; see “other software,” below.

Bee Docs Timeline 3D

Price: free-$25
Listing last updated: 1-2-2017
Available in the Mac App Store

Makes pretty timelines.

kSpectra

Version 3.9; Automator, Spotlight support;
Price: $100, available in the Mac App Store, Lite version $10; volume discounts
Listing updated 2-18-2020; software updated 2020

Published by Spectraworks, kSpectra Toolkit is “a set of programs for advanced spectral analysis of univariate or multivariate time series arising in many of the physical sciences, ranging from electrical engineering and physics to geophysics and oceanography, as well as biomedical sciences. The toolkit contains procedures for estimating the spectrum of a time series, decomposing the time series into trends, oscillatory components, and noise, and reconstructing the contributions of selected components of the time series,” according to the web site.

Dennis Kahlbaum wrote: “makes time series analysis relatively easy via GUI and built-in functions/methods.”

Typesetting and Graphs

LiveMath and PrintMath (formerly MathView, Theorist, MathEQ, Expressionist)

PrintMath 4.09, LiveMath 3.6
LiveMath: $129/year individual($69/year for students); $199 perpetual license; $249/year commercial
PrintMath: $39/year individual ($20/year students); $59 perpetual; $79/year commercial
Listing last updated: 10-9-2021

LiveMath and PrintMath, the latter for typesetting equations; only claims compatibility through Mac 10.14, “definitely will not work on Catalina” due to being 32-bit. Possibly in the process of being abandoned. Free trial for 30 days.

General math software

Statistical analysis (particularly time series analysis) can also be done using general-purpose mathematical software (especially with optional or standard plugins) such as:

Mathematica: Feature/version parity across Mac, Linux, Windows, and UNIX versions.
MATLAB: Gets considerable use for statistics, particularly Bayesian.
Maple: Windows, UNIX, and Linux also available.
IDL: OS X or UNIX native

Other software

ActivStats

Configurations available: MacOS X, Windows
Price: free
Listing updated 10-3-2021

Multimedia product designed to teach college-level statistics (with emphasis on DataDesk).

The site point sout that ActivStats Tools are based on the ActivStats statistics e-book, which includes t-tests, chi squares, and F tests; mean and regression boostraps; and randomization tests to do inferences. “These tools provide built-in datasets, but can also be used with any tab-delimited text data file for which the first row holds variable names.” Distributed by DataDesk.

gretl

Configurations available: OS X (Intel only)
Current version: 2016e
Price: free (open source)
Listing updated 1/2/2017

gretl is a cross-platform software package for econometric analysis, written in the C programming language. It is is free, open-source software that reads in numerous formats and can link to R. Features include an easy, intuitive interface, a wide variety of estimators, time series methods, output in tabular, equation, or LaTeX formats, a scripting language, command loop structure, and GUI for fine tuning graphs. (Thanks, Rudi Bekkers).

Insights (née Knowledge Miner (yX))

Configurations OS X 10.9 or newer, “optimized for Big Sur”
Current Version: 6
Price: Free basic version; ranges from Advanced ($129) to Ultimate ($1,990) with $35 student version and educational pricing
Listing updated 10-3-2021; software last updated Feburary 2021 to fix Big Sur bugs. Last major update, 2019.

Knowledge Miner’s data mining tool can now stand on its own, without requiring Excel. It can build predictive and descriptive models — “explicit and complexity minimized mathematical models,” with the ability to self-select inputs and show their importance. It has three languages — English, Spanish, and German.

Since Insights is now standalone, it’s able to run as full 64-bit software with parallel processing, boosting speed dramatically. Excel support is optional, for importing directly from Excel 2011 or 2016. Models can be exported in various formats.

The $129 Advanced version forecasts complex time processes without models, automatically runs simulations or forecasts real-time data (storing up to 5 steps ahead — the Ultimate version goes up to 120), exports models in generic form, can run up to 100 inputs with 50,000 samples, and supports up to 100 forecasting steps.

The Professional edition also does cost-sensitive modeling (including ROC and cost curve plots), also exports models in Objective-C and AppleScript, can run up to 2,000 inputs, with one million samples, and has unlimited forecasting steps. Ultimate adds self-organization modeling of interdependent, complex systems, can export systems of equation in Excel, and can handle 25,000 potential inputs.

The company’s own description:

INSIGHTS is original 64-bit parallel software for building predictive models from data, automatically, by evolutionary, self-organizing modelling approaches. Taking observational data that describes a problem, system, or process, the software constructs a working mathematical model... its AI-powered, self-organizing, modeling algorithms allow users to easily extract new and useful knowledge to support decision-making. ... Users in nearly any field can employ the easy-to-use software to analyze noisy data sets and build powerful models, which can be used to help to gain new insights into complex phenomena, predict future behavior, simulate "what-if" questions, and identify methods of controlling processes.

The software promises to hide the processes of model development, dimension reduction, variable selection, noise filtering, and model validation, self-organizing linear or nonlinear, static or dynamic regression models, to generate the equation that best describes the data. It also checks to see if the final model “reflects a valid relationship or if it just models noise.” It can create models for Status Quo or What-If problems, outputting models and model ensembles in ready-to-use Excel, Python, Matlab, AppleScript, or Objective-C code.

Data transfer / translation with Stat/Transfer

Version: 15
Price:
$499 (commercial),
$ 179 (academic subscription; $349 perpetual),
$ 99 (student one year subscription)
Listing last updated: 10-3-2021 (prices not checked)

Stat/Transfer can translate to and from most common statistical formats. It works surprisingly well, though you should check to make sure it can make the exact transitions you need. Version 15 brings support for Catalina and Stata 16.

Specialty software from Marley Watkins

This software is available from Marley Watkins at http://edpsychassociates.com/Watkins3.html (Thanks, Tricia Jones). Some of this software is PowerPC based and will not run under Lion; some are OS 9 based; other programs are Intel-based or universal binaries. There are some Excel spreadsheets, some FutureBASIC and BASIC programs as well.

MacParallel Analysis: tables of eigenvalues produced by Monte Carlo simulations.
Monte Carlo PCA for Parallel Analysis: computes Parallel Analysis criteria (eigenvalues) via Monte Carlo simulation.
r To z Conversion: calculates r-to-z and z-to-r.
Ability-Achievement Discrepancy Analysis: regression-based prediction of achievement for determination of a severe ability-achievement discrepancy.
Coefficient of Congruence: calculates the coefficient of congruence for comparing factor solutions from different studies or different samples.
Diagnostic Utilitity Statistics: calculates diagnostic efficiency statistics: sensitivity, specificity, false positive rate, false negative rate, positive and negative predictive power, odds ratio, Kappa, etc.
ROC Analysis: graphs the Receiver Operating Characteristic (ROC) curve and calculates the Area Under the ROC Curve (AUC) using a nonparametric method.
Chi-Square Analysis: calculates the chi-square statistic for 2x2 to 10x10 contingency tables. Also calculates the contingency, phi, and kappa statistics.
Probability Calc: calculates the probability of Z, t, F, and chi-square values.
MacBonferroni calculates Holm's sequential Bonferroni method to maintain error rates at a desired alpha level.
RandomNum generates random numbers within a range specified by the user.
UNITcalc calculates profile scores on the Universal Nonverbal Intelligence Test as per Wilhoit & McCallum (2002).
Simple t calculates independent and dependent t-tests from group data (mean, standard deviation, correlation).
MacKappa calculates general and partial kappa coefficients for nominal data.
CAcalculator-Revised calculates chronological age when given date of birth and date of test
rDiff calculates the reliability of difference scores, given the reliability of two tests and the correlation between then.
rProphecy calculates the Spearman-Brown 'Prophecy' formula that estimates the reliability of a test if its number of items is altered.
Effect Size Calculator calculates effect sizes (Cohen's d, r, Glass's É¢, Common Language Effect Size) given appropriate means and standard deviations.
MacOrtho transforms an hierarchical higher-order factor analysis solution into an orthogonal solution as described by Schmid and Leiman (1957).
MacMahalanobis calculates Mahalanobis distance for multiple cases, given entry of an inverted covariance matrix.
Pairwise Tests calculates three Pairwise Tests: simple Bonferroni adjustment, Holm's sequential Bonferroni method, and Benjamin and Hochberg's False Discovery Rate.
alphaN estimates the sample size needed for a specified coefficient alpha, given the Type I error rate and effect size.
rSim interactively simulates how the 95% confidence interval of the correlation coefficient and the statistical significance (p ≠ 0) vary as a function of sample size.
ASCA Assistant for the Adjustment Scales for Children and Adolescents (ASCA). It also performs syndromic profile analysis, discriminant analysis, and situtype analysis. Only for those who own ASCA manual and protocols.
Invariance calculates indices of factorial invariance: the coefficient of congruence, a chi-square goodness of fit test, and Cattell's Salient Variable Similarity index.