Analytics built for industrial data
Powerful multivariate methods, machine learning and design of experiments, with unique capabilities for spectroscopy and chemometrics.

Get it right from the start
Supports all steps from design to production, including advanced experimental designs with Design-Expert®.

Easy import of all types of data
Easy import of all types of data like material, sensor, process and spectral data from more than 30 different formats and new formats easily added.

Project based workflows
Work smarter and more efficient with project based workflows and functions that streamline all phases of the analytical process.

Unlimited new methods with Python
Python gives access to thousands of free scripts for use in Unscrambler with additional methods for import, preprocessing and machine learning.

Secure and compliant
Compliance mode, electronic signatures, user authentication and audit trails, and compliance with 21 CFR Part 11 and EU Annex 11.

Data import

• Generic import formats such as ASCII (text), MS Excel, Matlab, JCAMP-DX, NetCDF, JEOL, as well as generic database import
• Vendor specific formats from Thermo Fisher Scientific (GRAMS, OMNIC), Bruker (OPUS), Perten, rap-ID, Brimrose, ASD (Indico), Varian, Guided Wave (SpectrOn, Class-PA, NIRO JSON), FOSS (NSAS), PerkinElmer, DeltaNu, VisioTec and Viavi (MicroNIR™ Pro)
• Data and models from Design-Expert® and previous versions of Unscrambler can also be imported
• Some formats and database connections that are not listed above may be available as plugins. New formats easily added

Combining or reducing data

• Transpose
• Reduce (Average) along samples or variables
• Reshape using Row/Column major, Sequence wise or Level wise
• Augment or Append two or more matrices with matching dimensions
• Append two or more matrices based on column header names
• Flexible Sample Alignment by Polling, Event, Sample ID, Event within Sample ID
• Dimension Reduction for individual blocks of variables using PCA, PCR, PLSR

Scatter correction and other spectral transforms

• Smoothing with Moving average, Gaussian filter, Median filter, Savitzky-Golay
• Deresolve
• Normalization to common Mean, Max, Range, Area under the curve, Unit vector normalization, Peak normalization
• Baseline correction using Offset or Straight line
• De-trending
• Derivatives using Gap, Gap-Segment, Savitzky-Golay up to 4th order
• Standard Normal Variate (SNV)
• Multiplicative Scatter Correction (MSC)
• Extended Multiplicative Signal Correction (EMSC)
• Orthogonal Signal Correction (OSC)
• Correlation Optimization Warping (COW)

Descriptive statistics

• Missing values
• Level (Mean, Max, Min, Median, Quartiles)
• Range (Max-Min, Std., Variance, RMS)
• Distribution (Skewness, Kurtosis)
• Cross correlations
• Scatter effects

Statistical tests

• Equality of means (Paired t-test, Equal variance Student’s t-test, Unequal variance Student’s t-test)
• Equality of variances (F-test, Levene’s test, Bartlett’s test)
• Normality (Kolmogorov-Smirnov test, Mardia’s test of multivariate normality)
• Contingency analysis

Cluster analysis

• K-means, K-medians
• Hierarchical Cluster Analysis (HCA), including Single linkage, Complete linkage, Average linkage, Median linkage and Ward’s method

Explorative methods

• Principal Component Analysis (PCA)
• Rotated PCA (Varimax, Equimax, Quartimax, Parsimax)
• Multivariate Curve Resolution (MCR)

Regression methods

• Multiple Linear Regression (MLR)
• Principal Component Regression (PCR)
• Partial Least Squares Regression (PLSR)
• Support Vector Machines Regression (SVR)
• L-PLS Regression, incorporating three data tables

Classification methods

• Projection using PCA, PCR or PLSR models
• Soft Independent Modelling of Class Analogy (SIMCA)
• Linear Discriminant Analysis (LDA) with Linear, Quadratic, Mahalonobis options
• PCA-LDA, for classification of correlated data by LDA
• Support Vector Machines Classification (SVC)

Calibration transfer

• Interpolate
• Bias and Slope correction
• Piecewise Direct Standardization (PDS)

Spectroscopic transformations

• Absorbance to Reflectance/Transmittance
• Reflectance/Transmittance to Absorbance
• Reflectance to Kubelka-Munk
• Attenuated Total Reflectance (ATR) Correction

General and variance transforms

• Various Centre and Scale options
• Interaction and Square effects
• Weights
• Compute General, with operations such as log(x), 1/x, etc
• Quantile Normalize
• Fill missing
• Additive and Proportional Noise

Control charting

• Statistical Process Control (SPC) with Capability analysis
• Moving Block methods (Mean, Std., Relative std., F-test)

Input control

• Variable Limits filtering

Design of Experiments

• Two-level factorial screening designs
• General factorial studies
• Response surface methods (RSM)
• Mixture design techniques
• Combinations of process factors, mixture components, and categorical factors
• Design and analysis of split plots

Python scripting support

• Data import
• Preprocessing
• Machine learning
• Python resources

Batch modeling (plug-in – sold seperately)

• Modeling batch progression in relative time
• Prediction of new batch trajectories
• Any pretreatment of the data e.g. for spectra are stored within the model and applied for new batches
• The method is independent of sampling time, sampling period, batch progression and unequal batch lengths
• Dynamic limits for scores for individual components and the overall model
• Dynamic limits for the residual distance to the model (F-residual statistics)
• Contribution plot for drill-down functionality
• No missing value problem during prediction

Data import

Data handling

Diagnostics and plotting

Preprocessing and transforms

Building a PCA model

Interpreting a PCA model

Building a PLS model

Interpreting a PLS model