Based on your needs & wishes, I can suggest an appropriate training for you.
Or choose a theme:
Softeware usage: how does it work?
Reading data, writing data.
Creating new attributes (variables, fields, columns) from existing attributes
String (text attributes) manipulation
Join (match, merge) tables
Concatenate (add, union) tables
Charts and Crosstabulations
Data-mining: what predicts what?
There are two main goals to predictive modeling:
1. Learning the relationship between predictors and an outcome variable.
2. being able to predict for a new case the not (yet) known outcome variable, based on its relationship with other variables (predictors).
Supervised (predictive) learning techniques like Tree, Regression with Bagging or Boosting are used to predict the probability or value of events happening.
Unsupervised learning techniques like Cluster Analysis can be of specific help in this process.
Research design & Statistical tests: what causes what?
Experimental design allows us to make causal inferences.
Not essentially different from predictive modeling analyses, but different in data gathering method.
Univariate, Bivariate and Multivariate analysis. Basically these are all variations on the GLM (General Linear Model).