Learning Programming through Examples and Errors (for Non-programmers)

At this very moment, I only have notes for R. I will expand these to Python eventually.

What do you mean by teaching R via examples and errors?

The biggest challenge and annoyance for many non-programmers who want to learn programming is:

• The overwhelming details that create a steep learning curve yet seem useless for early programming
• The shallow amount of details when learning via copy/pasting code because you don’t have the confidence how things would change in practice.

This is written to strike a balance for people who feel both problems. To achieve this, we will introduce the programming concepts that immediately benefit some data analysis. We will also introduce examples and errors that will help correct your misconceptions of the code. We will constantly ask you to predict/guess what would happen under a different scenario to deepen your understanding.

Specifically, we will focus on data wrangling and simulations (e.g. working with non-rectangular data, joins, for-loops) but will not focus on data structures and algorithms efficiencies (e.g. sorting, recursion, differences between long vs double).

My assumptions

My assumption is that you have been exposed to

• basic statistical concepts like the average, sample standard deviation (vs population variance), and histograms.
• Used a calculator before

Computer setting up

I encourage you to set up Jupyter Notebooks on your computer so you could learn R, Python, or Julia in the future.

Problems

• Problem 1, How to simulate Law of Large Numbers?
  • Variables
  • Functions, arguments, default values
  • Random functions that create fake data
  • For-loops
• Problem 2, How to visualize data?
  • Data types beyond numeric values (character, enums/factors, booleans, data frames, missing values)
  • Subsetting data
  • Plotting
• Problem 3, Issues with naming variables
  • Scope
  • Using print() to understand the behavior of programs
• Problem 4, Data wrangling
  • Lists
  • JOINs between data tables
  • Wrangling with hierarchical (nested) data into rectangular formats
• Problem 5, Summarizing segments of data quickly
  • Vectorized functions and alternatives to for-loops
  • group-by then aggregate
• Problem 6, Working with text data
  • reading in raw text
  • Specifying patterns for computers (regular expression)
  • date/time values
• Problem 7, Scraping and calling APIs
  • Getting data from the internet programmatically
• Problem 8, Validating math using simulations
  • Hypothesis testing beyond the t-test: permutation tests
  • Intro to statistical computing, beyond data analysis!

What we won’t cover

What we won’t cover that you should learn in other computing courses:

• Data structure which is much deeper than the data types we’ve been discussing
• Algorithms, recursion, sorting which will deepen your knowledge of computational complexity
• Structuring code or a data pipeline
• Working backwards from a list of specifications.
• How to wrangle/analyze big data, i.e. where data is stored across multiple machines beyond just your computer.
• … many many more