Research

Overview

My work sits at the interface of applied mathematics, computational statistics, and data science.
I develop robust and interpretable methods for complex data analysis, with a special focus on reproducibility, open-source software, and practical translation into usable tools.

Research axes

1. Density estimation and anomaly detection

I develop methods based on GLBFP (Generalized Linear Blend Frequency Polygons) that unify histograms, ASH, and LBFP.
Goals: stability, adaptivity, and clear geometric interpretation.
Direct application: detection of atypical observations through combinations of local densities.
Software resource: GLBFP (R package, GitHub repository).

2. Directional statistics and angular models

Models for circular and directional data, including random circular effects.
Applications: orientation/movement, periodic phenomena, and dominant wind patterns.
Software resource: CircularRegression (R package, GitHub repository).

3. Hidden models and latent dynamics

Modeling with Hidden Markov Models for temporal and spatio-temporal data.
Recent interests: impact of temporal scale on inference, multi-individual models, and links with habitat-selection functions to connect behavior and environment.

4. Applied data science and pedagogical innovation

Transfer of statistical concepts to machine learning (measurement errors, interpretability, robustness).
Development of open resources for data science training (Quarto, learnr, GitHub) and reflection on responsible AI use in teaching.

Doctoral thesis

Title: General Multi-state Models for the Analysis of Animal Movement
Year: 2024
University: Université Laval
Supervisors: Thierry Duchesne and Louis-Paul Rivest

This thesis develops a general framework for multi-state modeling of animal movement, combining the rigor of hidden Markov models with the flexibility of habitat-selection functions.
It proposes a unifying formulation connecting behavioral and spatial approaches, along with a full software implementation in R.

Software and dissemination

  • GLBFP — Density estimation based on Generalized Linear Blend Frequency Polygons.
    A unifying method for histogram, ASH, and LBFP approaches, offering clear geometric interpretation and improved stability.

  • CircularRegression — Regression for circular and directional data.
    Includes extensions with random circular effects and three-dimensional models for movement in space.

  • site_ressources_SSD — Francophone educational resources platform for data science.
    Includes Quarto templates, interactive tutorials (Learnr), and reproducible examples for instructors and students.

  • GeneralOaxaca — R package available on CRAN for decomposition of gaps using the generalized Blinder-Oaxaca approach.
    It supports analysis of observed differences between groups (e.g., gender, regions, sectors) by decomposing contributions from composition effects and coefficient effects in linear and generalized models.

Publications

Published articles

  1. Nicosia, A. (2026). Discussion of “Addressing the Challenges of AI-Generated Assignment Submissions in Education: Insights and Strategies”. Journal of Data Science.
    https://doi.org/10.6339/26-JDS1208H — Article page

  2. Gagnon, S., Allard, M., Nicosia, A. (2017). Diurnal and seasonal variations of tundra CO₂ emissions in a polygonal peatland near Salluit, Nunavik, Canada. Arctic Science.
    https://doi.org/10.1139/AS-2016-0045

  3. Nicosia, A., Duchesne, T., Rivest, L.-P., Fortin, D. (2016). A Multi-State Conditional Logistic Regression Model for the Analysis of Animal Movement. Annals of Applied Statistics.
    https://doi.org/10.1214/17-AOAS1045

  4. Rivest, L.-P., Duchesne, T., Nicosia, A., Fortin, D. (2015). A General Angular Regression Model for the Analysis of Data on Animal Movement in Ecology. Journal of the Royal Statistical Society: Series C (Applied Statistics).
    https://doi.org/10.1111/rssc.12124

  5. Nicosia, A., Duchesne, T., Rivest, L.-P., Fortin, D. (2015). A General Hidden State Random Walk Model for Animal Movement. Computational Statistics & Data Analysis.
    https://doi.org/10.1016/j.csda.2016.07.009