Data Analysis and Visualization

class: center, middle, inverse, title-slide

.title[
# Data Analysis and Visualization
]
.subtitle[
## Chapter 3. Exploratory Data Analysis in R
]
.author[
### Iñaki Úcar
]
.institute[
### Department of Statistics | uc3m-Santander Big Data Institute
]
.institute[
### Bachelor in Data and Business Analytics
]
.date[
### <p>Licensed under Creative Commons Attribution <a href="https://creativecommons.org/licenses/by/4.0/">CC BY 4.0</a> Last generated: 2025-09-06</p>
]

---

class: base24, middle, clear

- [Types of Data](ch3.html#3)
- [Catalog of Graphs and Applications](ch3_1.html#2)

---
class: inverse, center, middle

# Types of Data

---
class: base24

# Forms of Data

### Entities

- The objects of interest, what we wish to visualize (e.g. a *sale*)
- Also known as **observations**, **records**, **instances**, **cases**...

### Relations

- Relationships between entities (e.g. a *customer* makes a *sale*)
- Many kinds: hierarchical, network, temporal, causal...
- Sometimes provided explicitly (e.g. edges in a network),
  sometimes its discovery is the purpose of the visualization (e.g. correlation)

### Attributes

- Properties of entities and relations (e.g. *price* and the *profit* of a sale)
- Also known as **features**, **variables**, **covariates**, **dimensions**...
- Can be qualitative or quantitative

---

# Types of Attributes

### Stevens' typology of measurement scales (1946)

- **Nominal**: purely qualitative, categories without order (e.g. fruits)
- **Ordinal**: qualitative with order (e.g. weekdays)
- **Interval**: quantitative data with no zero point (e.g. time of departure)
- **Ratio**: quantitative data with zero point (e.g. money)

### In practice

.font120[
- **Categorical data**: nominal or ordinal (e.g. fruits, weekdays)
- **Discrete data**: ordered categories or counts (e.g. number of visits)
- **Continuous data**: interval or ratio (e.g. time, money)
]

### Uncertainty

- Continuous data often come with a measure of uncertainty
- Can be represented as an additional attribute (e.g. confidence interval)
- Visualization of uncertainty is important but often overlooked

---
class: base24

# Data Analysis

.pull-left[
## Tidy Data

`\(N\times P\)` data frame:

|      | Attr1 | Attr2 | Attr3 | ... | AttrP |
|-----:|------:|------:|------:|----:|-------|
| Obs1 |       |       |       |     |       |
| Obs2 |       |       |       |     |       |
| Obs3 |       |       |       |     |       |
|  ... |       |       |       |     |       |
| ObsN |       |       |       |     |       |
]

.pull-right[
## Data Dimensions

- One dimension:<br>**univariate** analysis
- Two dimensions:<br>**bivariate** analysis
- More than two dimensions:<br>**multivariate** analysis
]