RPubs will retire in June 2027. Your existing documents will stay accessible through December 31, 2031
and Connect Cloud is the recommended home for new publishing. Read the blog post

Recently Published

Project 1 Clustering
Clustering
Association Rules for Business Trust Risk Signals (The InBillo Project)
This project applies association rule mining to a subset of the InBillo dataset in order to identify interpretable combinations of business characteristics associated with low customer trust. The analysis focused on non-score attributes such as firm age, size, legal form, financial transparency (debts) and online presence, using the Apriori algorithm to identify and extract repetitive patterns.
Global Development Patterns via PCA and Clustering
This report applies unsupervised learning methods to World Development Indicators (WDI) data to explore latent global development structures. Principal Component Analysis (PCA) is used to reduce dimensionality and identify interpretable development dimensions, followed by hierarchical clustering in the reduced space to derive stable country groups. The analysis emphasizes methodological justification, validation, and interpretability.
Reporting Flexdashboard by Prof. Dr. Solym Manou-Abi
Reporting Flexdashboard – Analyse des vols NYC 2013
HTML
users timeline
Reporting Flexdashboard – Analyse des vols NYC 2013
L'ensemble de données flights (associé au package nycflights13 en R ou utilisé dans des tutoriels Python/Pandas) est l'un des jeux de données les plus célèbres pour apprendre la science des données. Merci à notre best Professor Dr. Solym Manou-Abi.
Statistical model for newborn weight prediction
This project builds a statistical model to predict newborn birth weight using clinical data from 2,500 cases across three hospitals. Key predictors include gestational age, maternal smoking, infant sex, and biometric measures. The final regression model shows good performance (R² ≈ 0.79), addressing heteroscedasticity with log-transformation and incorporating interactions and non-linear effects. Diagnostic tests support model reliability. The model aids early identification of at-risk newborns, informs prenatal care (notably smoking cessation), and helps optimize neonatal resource planning. Limitations in extreme value predictions are noted, suggesting future validation and richer longitudinal data integration.