Kodemetrics

Want to make your analytical routines easier? Kodemetrics is the tool that enables exploratory and predictive statistical analysis and experimental design generation (DoE) without the need to know a specific programming language.

Programming languages are now essential for carrying out advanced experimental designs and analyses, but are often an entry barrier, becoming a significant waste of time for many domain experts. Kodemetrics helps researchers skip this barrier and focus on the research aspects of their studies.

Thanks to our experience in the chemometrics and research field, we have a clear understanding of experimentation needs and processes in labs. Indeed, we epitomized the most useful techniques of data analysis into a user-friendly tool, allowing users (whether they are students, researchers, labs or industrial R&D departments) to easily explore their experimental data. The selection of all the parameters required for each step of the analysis is guided in Kodemetrics through a simple graphical interface.

Exploratory Analysis: Principal Component Analysis – PCA

• Option to select the number of components to use

• Option to select variables to include

• Results visualization through scree plot, scores plot, loadings plot, T² and Q error plots with related contribution plots

• Model download

• Analysis report download in .docx format

Predictive Analysis: Linear Regression Models

• Preprocessing, model type and variable selection

• Coefficient estimates, p-values

• Diagnostic plots

• Response surface visualization

• Model download

• Analysis report download in .docx format

• Prediction of new observations

Predictive Analysis: Partial Least Squares Regression (PLS-R) and Discriminant Analysis (PLS-DA)

• Option to select the number of components to use

• Option to select variables to include

• Model summary with metric values (RMSE, R-squared, and MAE / Accuracy, Sensitivity, Specificity, Precision, Recall, etc.)

• Model goodness-of-fit plots

• Confusion matrix (for PLS-DA)

• Variable Importance plots and component coefficient plots

• Model download

• Analysis report download in .docx format

Mixture Models

• Preprocessing, model type and variable selection

• Coefficient estimates, p-values

• Diagnostic plots

• Response surface visualization

• Effects plot

• Model download

• Analysis report download in .docx format

Design of Experiments – DoE

• Full Factorial Design, with 2 or 3 factor levels

• Fractional Factorial Design, with 2-level factors

• Plackett-Burman Design

• Doehlert Design

• Central Composite Design (CCD)

• Box-Behnken Design

• Mixture designs with constraints on individual components and linear constraints between components

• D-optimal and I-optimal Designs

• Customization of factor levels and names

• Visualization of the experimental region

• Design download in .csv format

Multi-Criteria Decision Making – MCDM

• Pareto method, for choices based on a single criterion

• Simple Additive Ranking (SAR)

• Utility

• Desirability

• Dominance

• Multi-Attribute Utility Theory (MAUT)

• Technique for Order Preference by Similarity to Ideal Design (TOPSIS)

• Analytical Hierarchy Process (AHP), for decisions based on qualitative criteria

• Results in tabular and graphical format