Supervised and unsupervised methods. Multipole regression, logit model, discriminant analysis, classification trees. Principal component analysis, cluster analysis. Association rules.
The opportunities of big data analysis. Types of big data and their use.
Applications: market segmentation, credit risk assessment, consumer behavior, sentiment analysis (e.g. Twitter data). Analysis of logfiles.
Learning Objectives
Being able to prepare data and analysis for the main problems in business research.
How to use statistical software and interpret the results provided.
Prerequisites
Statistics, economic statistics (undergraduate)
Teaching Methods
Classroom lectures and computer labs.
Further information
Additional learning material will be available on the Moodle page of the Course of Economic Statistics (accessible from http://e-l.unifi.it/).
Type of Assessment
Written (computer session) and oral examination.
Course program
Introduction to business data analysis with reference also to the opportunities and limits offered by "big data". Types of research: exploratory, descriptive, explanatory. Primary and secondary data. Qualitative research vs. quantitative.
Supervised methods: multiple regression, dummy regressors, interactions between variables, non-linear forms; logit model; discriminant analysis; classification trees with categorical and non-categorical target.
Unsupervised methods: principal components, cluster analysis, association rules.
New data sources for market analysis: social media (twitter data), mobile phone data, and log file analysis, sentiment analysis. Laboratory using the R.