Advanced data analysis in R

Kauê de Sousa

This section is intended for users who want to work directly with tricot data in R and explore more advanced and customizable analytical approaches. While standard analyses can be performed through ClimMob or R-Instat, working in R allows full control over data processing, modelling, visualization, and reproducibility.

The following set of R packages has been developed to support different steps of the tricot data workflow, from data access to advanced statistical modelling.

climMobTools: Provides an API client to download and manage data directly from the ClimMob platform, making it easier to build reproducible analysis pipelines. https://cran.r-project.org/package=ClimMobTools

gosset: Offers methods for analysing experimental agricultural data, including tools particularly suited for tricot-style ranking data. It supports data preparation, model fitting, trait prioritization, data synthesis, and visualization. https://cran.r-project.org/package=gosset

climatrends: Supports the analysis of climate trends and agroclimatic indicators, which can be linked to on-farm trial data to better understand environmental drivers of performance. https://cran.r-project.org/package=climatrends

chirps: Provides access to CHIRPS and CHIRTS gridded climate data (rainfall and temperature), enabling integration of daily weather information into tricot analyses. https://cran.r-project.org/package=chirps

PlackettLuce: Implements the Plackett–Luce model in R, which forms the statistical backbone for analysing ranking data in the tricot approach. https://cran.r-project.org/package=PlackettLuce

Two introductory case studies are especially recommended.

Practical example with common beans for trait prioritization and crop performance

Practical example with gari/eba consumer testing

These examples illustrate how tricot data can be analysed in R to connect farmer evaluations, agronomic performance, and breeding decisions within a fully reproducible workflow.