Analytical chemist · Analytical measurement · Experimental design · Statistical modelling
I am a chemist working at the intersection of analytical chemistry, environmental measurements and quantitative data analysis.
My work focuses on turning experimental data into reproducible analyses, with particular attention to measurement uncertainty, method validation, calibration, experimental design and statistical modelling.
I use R and Quarto to make analyses transparent, reproducible and easier to review.
- Analytical and environmental chemistry
- Method validation and performance evaluation
- Calibration and analytical measurement
- Experimental design
- ANOVA, linear models and mixed-effects models
- Reproducible data analysis and reporting
R,data.tableandQuarto- Quality-oriented approaches to laboratory work
r-report_validazione
A reproducible framework for organising requirements, experimental data, calculations and reporting in analytical method validation under an ISO/IEC 17025-oriented workflow.
calibration_design
Simulation-based exploration of calibration design, heteroscedasticity, weighting, non-linearity and concentration prediction in chemical analysis.
analytical-anova
A progressive exploration of ANOVA, linear models, ANCOVA, interactions, diagnostics, robustness and mixed models using analytical-chemistry examples.
mixed_models
An experimental perspective on mixed-effects models, focusing on experimental units, non-independence, variance structure and uncertainty.
SIprecisa
R/Shiny application for evaluating precision and trueness of analytical measurements, with automated testing and deployment through Docker/ShinyProxy.
psa_pb_ny-alesund
Reproducible analysis supporting research on potential source areas of atmospheric lead reaching Ny-Ålesund.
I am interested less in applying statistical techniques mechanically and more in understanding how the experimental design, measurement process and scientific question determine the appropriate analysis.
The same principle guides my work on reproducibility: calculations, assumptions, data transformations and reporting should be explicit and, where possible, executable again.
R · Quarto · data.table · Git · GitHub Actions · renv
I also work with other tools when they are useful for the specific analytical or laboratory problem rather than as an end in themselves.
Most of the projects here are developed around real analytical questions or deliberately simplified experimental problems. The aim is not to demonstrate statistical techniques in isolation, but to understand when and why a particular model is appropriate.