Software for the Food and Beverage Industry

The Unscrambler® is the industry standard software product, used for chemometrics and process analysis in research and development, in the food and beverage industry. Tailor-made for modeling, prediction, and the classification of multivariable data sets, The Unscrambler® can also be used for the design of experiments.

A snapshot of industry applications

In the Food and Beverage industry, The Unscrambler® helps:

Case Studies

Some instances of how The Unscrambler® has been used in different Food & Beverage industry segments follow.

This needed to be ascertained from analytical measurements of moisture, pH, lactose, galactose, lactic acid, water activity, calcium and L-a-b appearance values.

The client needed to determine how to pre-treat and rearrange raw data, detect outliers, and build and test prediction models. The Unscrambler® was able to recommend the average shelf life of the whey. It determined y-response time using the intervals between the measurement and first fail date. PLS was used to detect outliers, select important chemical parameters and build prediction models.

The client employed sample measurements using a spectrophotometer, in order to measure moisture, pH, lactose, galactose, lactic acid, water activity, and calcium and L-a-b appearance values.

The client needed to determine how to pre-treat and rearrange raw sample data, detect outliers build and then test prediction models. Using The Unscrambler® Software, the client was able to accurately derive the optimal average shelf life of the formulation ingredient. It determined y-response time using the intervals between the measurement and first fail date. PLS regression modeling was used to detect outliers, select important chemical parameters and build prediction models.

The client specifically observed that in the wine industry where data has been collected over several years, the chemical composition of the grapes, relative to region and grower, is directly related to wine quality.

The client wished to identify the consistency in the rankings of the growers (i.e. good growers versus bad growers) over several years and whether the improvement of the scores of a particular vineyard would indicate an improvement in the quality of the grapes coming from that vineyard. The producer also needed to combine this information with chemical data, to investigate how chemical parameters correlate with Average Wine Quality scores and the relationship between chemical parameters and different growing regions.

Employing The Unscrambler®, the client generated grower rankings and consistency using Descriptive Statistic Analysis; evaluated tasters and vineyards with PCA (Principal Component Analysis); and estimated the effect of chemical parameters on wine quality using PLS regression coefficients and correlation loadings.

To produce decaffeinated coffee, manufacturers need to determine whether the percentage of caffeine is below or above 0.1 %. The traditional method for finding this is HPLC which is effective but expensive. The coffee company wanted to replace this method with one that is faster, easier and cheaper to use. Near Infrared Reflectance (NIR) spectroscopy fulfills these requirements.

The Unscrambler® helped the manufacturer classify samples and then develop a model that would predict the concentration of caffeine with high accuracy. Outliers are detected using the graphics features of The Unscrambler® and facilitate the easy interpretation of data. It is also possible to determine which samples and wavelengths are the most important for the model. From the results obtained, it was easy to optimize the model to result in the best prediction ability.

 

 

Resources Methods Case Studies Application Notes Suggested Readings