CougarStats: Integrating Computational Statistics and Interactive Visual Analytics

A. Krishnamurthy${}^a$, M. Abou El Nasr ${}^a$ and J Bennett ${}^a$

${}^a$Mount Royal University

CougarStats (www.cougarstats.ca) is an open-source web-based platform developed in the R ecosystem that integrates statistical modelling and visualization within a unified environment for data analysis.

The platform provides access to core statistical techniques through interactive modules spanning descriptive statistics, statistical inference, regression modelling, and machine learning. Rather than treating modelling and visualization as separate steps, CougarStats supports interactive analytical workflows in which statistical estimation, diagnostic assessment, and graphical exploration occur simultaneously. This integration reduces the cognitive gap between computing a result and interpreting its significance.

A central design principle of CougarStats is computational transparency in statistical computing and visualization. Analytical workflows are implemented using established R methods while exposing step-by-step calculations and dynamically linked graphical output that facilitate interpretation of model behaviour and underlying data structure.

The platform continues to expand to support a broad range of modern data science workflows, including supervised classification methods (e.g. Linear Discriminant Analysis (LDA)) and unsupervised dimensionality reduction techniques (e.g. Principal Component Analysis (PCA)), and predictive modelling for time-series data.

Keywords: Data science, Computational Statistics, Visualization.