Course Introduction

# 00 - Course Introduction

## Data Science with R &#183; Summer 2021

### Uli Niemann &#183; Knowledge Management & Discovery Lab

#### [https://brain.cs.uni-magdeburg.de/kmd/DataSciR/](https://brain.cs.uni-magdeburg.de/kmd/DataSciR/)

---

&nbsp;

## Welcome to<br />Data Science with R!

]

]

???

- course organization + first tour of R and RStudio

---

class: bottom
background-image: url("figures/00-stay-at-home.jpg")
background-size: fill

---

## Who are <u>YOU</u>?

- You are studying either **DKE** or another Master FIN degree with focus on data science.
- You have heard one or more **introductory data science courses** in the previous semesters.
  - "Data Mining I" by Prof. Spiliopoulou
  - "Machine Learning" by Prof. Nürnberger
  - "Visual Analytics" by Prof. Preim
  - ...
- While you liked these courses to learn fundamental data science concepts, you 
have not had the chance to apply this theoretical knowledge in a **practical 
project on real-world datasets**.
- You are most proficient in **Python** (or **Java**).
- You have no or very **little experience with R**, but are keen to learn it either 
because you either heard of R's reputation to be a powerful language for 
data science projects or simply to complete your programming skill set.

]

]

---

## Data Science with R

???

- at the end of the semester, you should be confidently able to use R to gain insights from data 
- workflow of a typical DS project
- import data from a file, database, website or web API and load as data frame into R
- transform and clean
  - bring it into a form that is consistent with the semantics of the data
  - transform: feature engineering: derive new variables or do some feature selection or reduction (required most of the time)
- visualization: shows things that you did not expect, or raise new questions about the data
  - hints that you are asking the wrong question or need to collect different data
- models: classification, regression
- communication: crucial to communicate to others
- DS often involves a lot of trial and error
  - don't know the best approach in advance
- often starts with some initial questions
  - initial questions are often quite vague, because you are working on an unsolved problem
  - improve quality of the questions after EDA (better data understanding) or first modeling results

https://simplystatistics.org/2019/04/17/tukey-design-thinking-and-better-questions/

---

## Course structure

]

]

???

Part 1: until first week of June

- asynchronous content: slides, videos, exercises, and quizzes
- synchronous elements: weekly meetings (code-along, questions)

Part 2: throughout the entire semester

---

## Website & syllabus

???

first part: in each week we cover a different topic 
- videos
- slides
- exercises
- quizzes
- additional reading

- crash course is not meant to be exhaustive
- "common ground" like packages that all of you are most likely going to use in your projects 
- however, you are encouraged to use also other packages

After successful completion of this course, you will be able to proficiently perform the following tasks in `R`:

- **import and preprocess raw data**
- **transform data** for modelling
- perform **exploratory data analysis** with **summary statistics** and **visualization**
- understand, build and evaluate predictive **classification and prediction models**, including **regression** models, **tree-based** models, **ensembles** and **boosted** models
- communicate and disseminate results and findings through **reproducible documents**, presentations, websites and **interactive web applications**

---

## Resources

]

]

???

- Website: hub for all materials and news about the course
- GitHub: course materials are hosted there (slides, exercises + data)
- Moodle: video recordings, quizzes, forum
- Zoom: during the first part of the course - weekly meetings with code-alongs + Q&A

---

## Quizzes

- Handled on E-Learning (Moodle)
- 7 quizzes in total with 5 questions each
- Each quiz open on Mondays 00:00 and is due on the following Sunday at 23:59 German time
  - The first quiz is available between 12.04 00:00 and 18.04 23:59
- Work on them individually
- Lowest score is dropped, i.e., only your 6 best results count
- Total score of correctly answered questions gives you up to 30 pt for your final grade
- Questions must be answered in a fixed order, which means you **cannot** return to previous questions

???

- individual part:
- purpose: motivate you to engage with the crash course part of the course

---

class: middle
background-image: url("figures/00-example-questions.png")
background-size: contain

---

## Project

In a nutshell: **_Find an interesting dataset and do something with it!_**

- analyze a dataset of your choice
- demonstrate your proficiency in the techniques we have covered in this course (+ desirably beyond)
- **quality over quantity**: don't compute every possible statistic, but :
    - show that you can **ask meaningful questions and answer them using R**
    - display that you are **skilled at interpreting and presenting your results**
    - critically scrutinize your approach (**adequacy**) and your results (**reliability, validity**) and make suggestions for improving your analysis
- emphasize on **compelling visualizations** and a **convincing, coherent and clear narrative**
- all analyses must be done in R; you are encouraged to use tidyverse packages but you don't have to

???

- team work part
- Moodle forum to form teams

---

## The project data

- choose any dataset which is _not analyzed to death_ (no Kaggle, UCI ML repo, or similar)
- the dataset may already exist or you may collect your own data via a survey or an experiment
- the dataset may be based on your personal interests or based previous projects in other courses

---

???

- team size: 3-4 ideal
- show example projects

- description
- project team
- milestones
- team formation
- topic
- R Markdown process notebook
- Code
- project website
- screencast
  - Please ensure a sufficient sound quality.  
  - It may be worthwhile to borrow or invest in an external USB microphone.
  - What do you feel is the best part of your project?
  - What insights did you gain?
  - What is the single most important thing you would like your audience to take away?
  - Make sure it is upfront and center rather than at the end.
- final presentation
- submission instructions
- hall of fame

---

## Deliverables & Hard Deadlines

&nbsp;

Date          | Description
:------------ | :---------------------------------------------------------
20.05. ~~16.05.~~        | Team formation & project proposal submission; registration
24.-28.05.    | Project proposal feedback
06.07.        | Final project submission due
09.07.        | Final presentations, exact time and location to be announced

]

]
.pull-right40[

]

---

## Your final score in the course

Your overall course grade will be determined by the following components:

1\. **30 pt**: Weekly quizzes  
2\. **70 pt**: Project:

- 8 pt: Project proposal
- 25 pt: Quality of R Markdown notebook with respect to correctness, comprehensibility and reproducibility
- 10 pt: Complexity and level of difficulty of the project
- 5 pt: Completeness and overall functionality of the repository and website
- 8 pt: Screencast
- 14 pt: Final presentation

---

## Zoom meetings

We will have weekly Zoom meetings. They are completely optional.

The Zoom meetings are mainly for:

* code-alongs
* Q&A

Expectations for Zoom meetings:

In larger sessions you should

- have your microphone muted by default
- use the _raise hand_ feature or type questions and comments in the chat

In small team sessions you should

- have your camera turned on as much as possible
- engage with your team mates via voice and text chat
- take turns sharing your screens when necessary

---

## Collaboration regulations

- Only work in the context of the team project should be completed collaboratively.
- Exercises should be completed individually.
- Quizzes must be completed individually.
- You are welcome to discuss problems in general and ask for advice.

---

## Sharing and reusing code policy

- You can use any online resources, but you must explicitly state where you got the code that you use directly or use as inspiration for your solutions.
- Any reused code that is not explicitly cited will be treated as plagiarism, regardless of the source.

---

## Late work policy

- Submissions after the deadline will not be considered.
- General advice: Don't wait until the last minute.
- But: Email me if there are reasons beyond your control that prevent timely submission.

---

# Thank you! Questions?

&nbsp;