El CoDaCourse’2016 es un curso sobre el análisis de Datos Composicionales que se impartirá en la Universidad de Girona del 4 al 8 de julio, 2016. Está organizado por el grupo de investigación en Estadística y análisis de Datos Composicionales. Este curso está oficialmente acreditado por la International Association for Mathematical Geosciences (IAMG). Más información en http://www.compositionaldata.com/

Objectives and contents

Compositional data are vectors which components show the relative importance of some parts of a whole. Typical examples are data presented in percentages, ppm, ppb, or the like. Aitchison introduced thelogratio approach to analyse compositional data back in the eighties. Since then, progress has been done in understanding the geometry peculiar to their sample space, the D-part simplex. 

This CoDaCourse provides an introduction to theoretical and practical aspects of the statistical analysis of compositional data. The following topics will be covered:

  • The sample space, principles of CoDa.
  • The Aitchison geometry on the simplex.
  • Coordinate representation; distributions on the simplex.
  • Exploratory analysis (centering, variation array, biplot,balance-dendrogram).
  • Irregular data: zero values, outliers and missing data.
  • Introduction to multivariate analysis: regression, manova,cluster and discriminant.

Duration and language

One week 4-8 July 2016

5 days, 25h class room

Language: English

Teaching staff

The teaching staff will be composed by members of the research group on Compositional Data Analysis that includes professors from the University of Girona (UdG) and form the Technical University of Catalonia (UPC).

Teaching methods

The course will consist of theoretical and practical sessions, a hands-on session as well as case-study presentations followed by an open discussion session. We also are working to organize a session with aninvited teacher.

In the theoretical sessions, the current state of the art in this field is presented. In the practical sessions, coda techniques are applied using the freeware software CoDaPack and some R packages as zCompositions for the imputation of zeros. CoDaPack is freeware developed for the statistical analysis of compositional data (http://imae.udg.edu/codapack/).

The open discussion session is a case-based discussion session. Some compositional data sets and their particular problems will be presented, discussed and analysed interactively. Assistants to the course are encouraged to bring their own data sets and state those questions they would like to be answered during the course. Some of these proposals will be selected for a detailed open discussion.

For this edition we have a new proposal, a hands-on session. With some outputs and specific questions related to a real data set, the participants will be invited to discuss, choose and build a reasoning applying Compositional Data methodology.

Preliminary Schedule

9:00-13:00 14:30-17:00
July,4 Sample space, principles, log-ratios, geometry  
July,5 Exploratory (centering, variation array, biplots) Coordinates, balance- dendrogram, distributions
July,6 Irregular data (zeros, missing data, outliers)  
July,7 Invited teacher: Robust compositional data (to be confirmed) Multivariate methods (manova, cluster), linear processes
July,8 Open discussion, hands-on session  

Target group

Statisticians and applied scientists of any field, in particular engineers, geologists, environmental scientists, business statisticians, sociologists, economists or biologists, working for academic or industrial institutions. It is strongly recommended that attendants have undergone some first semester courses on statistics, algebra and calculus. Basic knowledge about multivariate statistics may also be handy.

How to apply Register on the website www.compositionaldata.com

and fill in the form at the

CoDaCourses menu

Course fees

Regular fee: 250€

Student fee: 150€ (proof of student condition is required)

Late fee (after June 1, 2016): 350€

The fee will included course material, coffee breaks and lunches (Tuesday, Thursday).