Course Overview

.title[
# Course Overview
]
.subtitle[
## Introduction to Data Science (BIOL7800)
<a href="https://introdatasci.dlilab.com/" class="uri">https://introdatasci.dlilab.com/</a>
]
.author[
### Daijiang Li
]
.institute[
### LSU
]
.date[
### 2023/08/21
]

---

### Assistant Professor 
 
### Department of Biological Sciences

### Center for Computation & Technology

### https://www.dlilab.com
]

---

# My role

- Help you learn these materials

- Help you learn how to use what we learned in your research

- Help you learn how to ask for help

- Be a potential future resource
]

---

# My role

---

# Who are you?

???

Introduce yourself: department, lab, research direction/interest, etc.

---

# Go Through the [Syllabus](https://introdatasci.dlilab.com/syllabus/)

---

# How are we going to go through this course?

### Learning by doing (through trials and errors)

### Use lots of online resources

### Peer-teaching and learning; collaborative coding

### Google and Stack Overflow

---

# What is _data science_ ?

---

# Data science is interdisciplanary

---
class: middle, inverse

.font200[
> Data Scientist = statistician + programmer + coach + storyteller + artist  
>
> .right[Shlomo Aragmon]
]

.font200[
> Good data science is distinguished from bad data science primarily by a repeatable, thoughtful, skeptical application of an analytic process to data in order to arrive at supportable conclusions. 
>
> .right[Jeff Leek]

]

### Applies to every discipline

---

# The age of big data

.font150[
> Between the dawn of civilization and 2003, we only created five exabytes<sup>1</sup> of information; now we’re creating that amount every two days. 
>
> .right[Eric Schmidt et al, Google]

]

???

ask students what kind of large datasets exist in their field of research

---

# Data science processes

2. Get the data

3. Clean and prepare the data

4. Explore the data

5. Fit models to extract insights

6. Tell, explain, and illustrate results
]

---

# Data science processes

2. .red[Get the data]

3. .red[Clean and prepare the data]

4. Explore the data

5. Fit models to extract insights

6. .red[Tell, explain, and illustrate results]
]

???

These steps are the most time consuming ones, so better to make them (and others) reproducible

---

![data science workflow](figs/workflow.png)
.right[ Blitzstein & Pfister, 2015]

---

# Questions in data science

---

background-image: url('figs/q_types.png')
background-position: 50% 50%
background-size: contain

---

# Examples

| Types                     | Examples                                                                         |
|---------------------------|----------------------------------------------------------------------------------|
| Descriptive               | Proportion of different races in the USA                                         |
| Exploratory               | Investigate correlations among multiple variables                                |
| Inferential (most common) | Does air pollution correlate with life expectancy at the state level in the USA? |
| Predictive                | Using polling data to predict election results; not necessarily explain why      |
| Causal                    | *Average* risk of COVID for vaccination vs non-vaccination                       |
| Mechanistic               | Impacts of wing design on air flow over a wing; rare outside of engineering      |