What You Should Know Now — and What's Next

What You Should Know Now

• Core syntax: assignment, vectors, types, missing values, conditionals, loops, and functions
• Lists, factors, matrices, and data frames
• Importing, cleaning, reshaping, and transforming data with the tidyverse
• Visualising data with ggplot2
• Descriptive statistics, hypothesis tests, and regression models
• Functional programming with the apply family and purrr
• Error handling, reproducible reports with Quarto, and testing with testthat
• Machine learning workflows with tidymodels
• Environments, S3/S4/R6 objects, performance tuning, and databases
• Shiny web apps, APIs with plumber, and building your own packages

What Should You Learn Next?

• Statistics in depth — study design, generalized linear models, mixed models, Bayesian methods, and causal inference. R's strength is statistics; the language is the easy part.
• SQL — most real-world data lives in databases, and SQL is the language for getting it out.
• A related language — Python, the other major data-science language, especially for deep learning and production machine-learning systems; many teams use both.
• Communication — Quarto for reports, books, and websites, and good chart design, so your analysis actually changes decisions.
• Specialized fields — Bioconductor for bioinformatics, sf and terra for geospatial data, tidytext for text mining, and fable for forecasting.

Build Something

Pick a question you genuinely care about — in your job, your hobbies, your city — find data that could answer it, and publish the analysis as a Quarto report or a small Shiny app. A finished analysis of real data is the best portfolio piece a data scientist can have.