Benchmarking Kendall's Tau in R and Rcpp
Implement and benchmark a fast Kendall’s tau-a in C++ via Rcpp against base R, discuss tie handling (tau-b), and when to move from R to C++.
About Me
I am Thiago de Paula Oliveira, a statistician at AbacusBio with 14+ years of experience turning experimental, genomic, and performance data into decisions. I specialise in advanced mixed-model and Bayesian analytics, the development of economic and sustainability selection indexes, and the delivery of reproducible analytics products—from R/C++ codebases to Dockerised dashboards. Whether the brief is accelerating genetic gain, improving farm-system resilience, or supporting athlete health, I focus on rigour, transparency, and decision-ready outputs.
Interests
- Statistics and biostatistics
- Concordance analysis
- Multilevel and forecast models
- Generalized mixed-effects models
- Longitudinal data
- Quantitative genetics & breeding analytics
- Agricultural decision-support dashboards
- Reproducible analytics pipelines for agri-genomics
- Economic and sustainability selection indexes
Education
PhD in Statistics
University of São Paulo
MSc in Statistics
University of São Paulo
BSc in Agricultural Engineering
University of São Paulo
📈 Expertise and Research
I am a statistician with 14+ years of experience turning noisy experimental, genomic, and performance data into decisions. After completing my PhD in Statistics at the University of São Paulo, I specialised in advanced mixed-model/Bayesian analytics and in the development of economic and sustainability selection indices that keep breeding programmes accountable.
As a Consultant Statistician at AbacusBio, I lead cross-functional teams that deliver genetic-evaluation pipelines, automated QC/ETL workflows, and decision dashboards for livestock, crop, and agri-tech partners. That work depends on production-grade code in R/C++/Bash, Docker-based reproducible environments, and early collaboration between domain scientists and data engineers.
Earlier, I held a Marie Skłodowska-Curie COFUND fellowship at the Roslin Institute (University of Edinburgh), built predictive health and sports-analytics products at the Insight Centre (NUI Galway) and Orreco, and lectured in statistics at USP. Along the way I have published across Nature-branded journals, advised national breeding programmes, and mentored teams on delivering transparent, auditable analyses.
Whether the brief is accelerating genetic gain, improving farm-system resilience, or supporting athlete health, my bias is toward rigour, reproducibility, and decision-ready outputs. Browse my recent publications and projects, and get in touch if you would like to collaborate or have a specific challenge in mind.
Featured Publications
Recent Publications
(2025).
Breeding for sustainability: Development of an index to reduce greenhouse gas in dairy cattle.
Animal.
(2024).
A Hierarchical Approach for Evaluating Athlete Performance with an Application in Elite Basketball.
Scientific Reports.
(2023).
Developing best practices for genotyping-by-sequencing analysis in the construction of linkage maps.
GigaScience.
(2023).
Comparison of Markerless and Marker-based Motion Capture Systems using 95% Functional Limits of Agreement in a Linear Mixed-Effects Modelling Framework.
Scientific Reports.
(2023).
Pedigree-based Animal Models Using Directed Acyclic Graphs.
Under consideration in Livestock Science.
Cited by
Recent & Upcoming Talks
Recent Posts
From prototype to production, choosing the right R and C++ tool in Rcpp
what each tool solves, when to reach for it, and ready-to-paste code.
Effects of compression techniques on data read/write performance
This post explores the performance of various compression techniques in R for reading and writing operations, highlighting file size, speed, and memory usage.
Comparing data read and write performance in R
This post explores the performance of various data formats in R for reading and writing operations, highlighting file size, speed, and memory usage.
Exploring polynomial, fractional polynomial, and spline models
1 Polynomial models 2 Fractional polynomial models 2.1 Finding optimal power Values in fractional polynomials 3 Spline models 3.1 Example 3.2 Challenges 3.3 Selection process for spline models 4 Citation The ability to accurately model and interpret complex data sets is paramount. This technical exploration delves into three sophisticated modelling techniques: - Polynomial Models, - Fractional Polynomials, and - Spline Models. Each of these models serves as a fundamental tool in the statistical toolkit, enabling us to capture and understand the intricacies of linear and non-linear relationships inherent in real-world data.
🎓 Connect with an expert statistician
I focus on advanced statistical modelling, economic and sustainability selection indices, interactive dashboards, and reproducible (Dockerised) pipelines that deliver decision-ready insights for agriculture, genetics, and sports performance.
Areas of impact
Agriculture. Design and analyse agronomic and farm-systems experiments, including multi-environment trials and spatial models, to optimise yield, resource use, and sustainability.
Genetics. Build genetic-evaluation pipelines and economic and sustainability selection indices that maximise genetic gain and inform breeding objectives.
Sports analytics. Develop tools and applications that enhance athlete performance through data-driven insights.
Explore my publications, projects, and recent work. If you are interested in collaborating or would like to learn more, please get in touch.
Stay connected and follow my work in statistical modelling and data analysis: