Data Science / Statistics / R Foundation / R / Exploratory Data Analysis using R
| ID | 2858624 |
|---|---|
| Classroom | 3 days 9:00-16:30 |
| Webinar | 5 days 9:00-12:30 |
| Method | Presentation with examples and hands-on labs. |
| Prequisite | Basics in R and Statistics |
| Audience | Data Analysts |
Exploratory Data Analysis (EDA) is a statistical approach to analyzing data sets to summarize their main characteristics. This training primarily focuses on four main techniques of EDA: Principal Component Analysis (PCA) for quantitative variables, Correspondence Analysis (CA) and Multiple Correspondence Analysis (MCA) for categorical variables and finally (hierarchical and partitioning) clustering methods. As an umbrealla technique, this training also shows Factor Analysis (FA) and Multiple Factor Analysis (MFA). For the hands-on labs and practical examples the participants will use R and esp. FactoMineR - a special R package for the exploratory data analysis.
Objectives of PCA and Introduction to PCA - Studying Individuals: The Cloud of Individuals, Fitting the Cloud of Individuals - Variables: The Cloud of Variables, Fitting the Cloud of Variables - Relationships - Interpreting the Data - Testing the Significance of the Components - Implementation with R and FactoMineR
Objectives and the Independence Model - Fitting the Clouds: Row and Column Profiles - Interpreting the Data - Implementation with R and FactoMineR
Objectives: Studying Individuals, Variables, and Categories - Defining Distances between Individuals and Distances between Categories - CA on the Indicator Matrix: Relationship between MCA and CA, The Cloud of Individuals, Variables, and Categories - Implementation with R and FactoMineR
Concepts of Similarity and Distance: Similarity between Individuals and Groups - Ward's Method - Partitioning and Hierarchical Clustering - Direct Search for Partitions: K-means Algorithm - Clustering and Principal Component Methods - Implementation with R and FactoMineR
Factorial Analysis of Mixed Data - Weighting Groups of Variables - Comparing Groups of Variables and Indscal Model - Qualitative and Mixed Data - Multiple Factor Analysis and Procrustes Analysis - Hierarchical Multiple Factor Analysis - Implementation with R and FactoMineR
Marco Skulschus (born in Germany in 1978) studied economics in Wuppertal (Germany) and Paris (France) and wrote his master´s thesis about semantic data modeling. He started working as a lecturer and consultant in 2002.
- He works as an IT-consultant and project manager. He developed various Business Intelligence systems for industry clients and the public sector. For several years now, he is responsible for a BI-team in India which is mainly involved in BI and OLAP projects, reporting systems as well as statistical analysis and Data Mining.
He led several research projects and was leading scientist and project manager of a publicly funded project about interactive questionnaires and online surveys.
