Exploratory Data Analysis & Reporting

Gather, visualize and analyze data to gain insight and make fast, informed decisions. Use Material Science Predictive Analytics for comprehensive, collaborative modeling and simulation applications for Material Sciences discovery research.

When formulating a new food recipe for example there are multi-dimensional measurable characteristics that need to be taken into consideration, which include viscosity, pH factor and nutritional parameters.  Some companies are going beyond physical attributes and creating sensory profiles to predict sensory properties.

Sensory models detect if there's a significant, perceptible difference between products when a change is made to one or many dimensions.  If a difference is identified, a descriptive analysis is performed to determine how the products compare across different sensory characteristics. The food scientist then analyzes individual texture attributes, such as mouth coating, viscosity and astringency to hone in on which attribute causes it to be different from the others. These models can provide a comprehensive set of sensory terms to standardize how product texture is described during the development process.

Using in silico models can optimize new recipes by selectively focusing testing on the candidates to most likely to succeed.   The most common scenario in CPG business is reformulating to fit a set of new conditions, typically by substituting an active ingredient, due to cost or shortage of product availability.  Reformulation needs to consider the cost, but also the performance of product which must remain at a comparable level. Cost, performance and environmental impact are just some of the factors that can be driving the model.

Benefits of the Exploratory Data Analysis & Reporting solution include:

Reducing the cost and time associated with physical testing and experimentation through “Virtual Screening” of candidate recipe variations.

Accelerating the innovation process - developing new, better performing, cost effective and more sustainable recipes faster than can be done with physical processes.

Using powerful informatics capabilities to optimize the recipe based upon the desired dimensional constraints.