Data Analysis and Modelling

We have expertise in data analysis applied to ecology, encompassing population dynamics (e.g., count data, capture-mark-recapture), eco-epidemiology (e.g., SEIR models, survival analysis), and ecotoxicology (e.g., dose-response, toxicokinetic-toxicodynamic).

We employ various statistical techniques ranging from classical analyses (hypothesis tests, additive or generalized linear mixed-effects models) to more complex hierarchical models using Bayesian inference.

To give you an idea of our capabilities, you can freely explore (Open Source licenses) on our gitlab repository.

Our Lectures

A set of Jupyter Notebooks, which we can deploy for you in a JupyterHub, covering various domains of ecological modeling (differential equations, SEIR models, sensitivity analysis, dispersion models), and an Advanced Ecotoxicological Modelling course introducing Bayesian Inference (LM to GLMM), TK-TD models and their inference, including GUTS (Generalized Unified Threshold models of Survival) survival models, and DEB (Dynamic Energy Budget) theory

Our Softwares

Developed in partnership with the MEPS laboratory of the University of Lyon 1. The R packages include:

morse: Bayesian Inference and Analysis of survival and reproduction data.
rPBK: Bayesian Inference of PB-TK (Physiologically Based ToxicoKinetics) models.
rbioacc: Bayesian Inference of multi-compartment, multi-compound bioaccumulation models.

Our Web-Tools

BAPPviewer: Web-interface for graphic representation of soil-plants transfert datasets BAPPET and BAPPOP.

Collect, Analysis, Interpretation

Complete management of your data analysis workflow, from sampling plan to statistical analysis choices and interpretation.

Training for Statistical Analysis

In person or through access to servers we provide, we can train you on all or part of the mentioned tools. Learn more

We help you gain a thorough understanding of ecosystems by accurately interpreting the results of appropriately collected data analyses.

We assist you in these key aspects (collection, analysis, interpretation) either by working with you on the problem or by teaching you the necessary skills to successfully learn about and manage ecological and agro-ecological systems.

Risk Scenario Modeling

We have recognized expertise in scenario modeling in eco-epidemiology and ecotoxicology (PBK, TK-TD, DEB), as well as in multi-agent models.

These projects are custom-made and have resulted in research publications:

Data Analysis

Data is the foundation upon which final conclusions (like on Risk Assessment) are based, so it is essential to collect data that is suitable for your intended use and analysis. We can help you:

Determine the type of data that needs to be collected
Determine the sample size
Optimally allocate survey efforts
Evaluate possible field methods
As we specialize in ecological applications, we are familiar with many of the specific challenges and practical limitations you often face.

We aim to provide you with high-quality data analyses in an accessible language, which you can then communicate in concrete terms for the client's problem:

Mark-Recapture
Distance Sampling
Occupancy Modeling
Species Distribution Modeling
Spatio-Temporal Modeling
Generalized Linear Models
Generalized Additive Models
Linear Regression
ANOVA, t-tests, z-tests, etc.

Machine Learning

The term Machine Learning (machine learning) today encompasses the entire spectrum of data analysis, from hypothesis testing to the sophisticated neural networks of language models. We have already employed neural networks for satellite image recognition (convolutional neural networks, CNN) as well as direct-action neural networks in differential equations to account for mixtures of contaminants in exposure-uptake-effects models.