Chemometrics web app's part 2: Dimensionality reduction and exploratory analysis

• Published in: Chemometrics and intelligent laboratory systems Vol. 237; p. 104810
• Main Authors: Darzé, Bernardo Cardeal, Lima, Igor C.A., Luna, Aderval S., Pinto, Licarion
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 15.06.2023
• Subjects: Data test; Non-parametric; Pattern recognition; PCA; RShiny; Data test; RShiny; Non-parametric; Pattern recognition; PCA
• ISSN: 0169-7439; 1873-3239
• DOI: 10.1016/j.chemolab.2023.104810

This work reports the release and the usability of the dimensionality reduction app, an R application developed with the RShiny package to reduce data dimensionality using filters, resolution reduction, and algorithms based on parametric and non-parametric approaches. This application accepts data directly from.txt,.csv,.xlsx file extensions, and R data files directly exported by the data handling app presented in the Chemometrics web app's part 1: Data handling manuscript. The idea is to use the data directly from the instrument or preprocess it with the first app. Therefore, using the application does not require a deep knowledge of R software or matrix manipulation. The dimensionality reduction app allows performing basic Preprocessing such as mean centering and autoscaling, some adequacy testing, some filtering methods to reduce the data dimensionality, cluster analysis, diagnosis of the number of clusters, parametric and non-parametric based algorithms for dimensionality reduction and exploratory analysis, such as PCA and T-SNE respectively. Three datasets will be used for data dimensionality reduction and exploratory analysis. All the figures presented in this manuscript were directly exported from the dimensionality reduction app to highlight its functionalities. The main idea of this application is to allow chemometrics users to perform several strategies for dimensionality reduction and exploratory analysis even with no knowledge of programming language. Besides, the dimensionality reduction app is an open-access code available to be used on RStudio free environment at the software or cloud computing using computers, tablets, smartphones, or similar devices, even without installing anything for usage. •Open-source web app developed in R for dimensionality reduction and exploratory analysis.•Data adequacy test, variable preprocessing, filtering methods and simple resolution reduction.•Cluster Analysis and diagnosis tools to evaluate the number of clusters.•Parametric methods for dimensionality reduction and exploratory analysis.•Non-Parametric methods for dimensionality reduction and exploratory analysis.