R & RStudio for Marine Scientists

R was written by statisticians, which is why GLMs, mixed models and multivariate methods are central to it rather than add-ons. Its marine packages cover the field from CTD casts to biodiversity records: oce, marmap, heatwaveR, vegan, robis and rerddap. A script records every step, so peer reviewers and regulatory auditors can follow exactly how a number was produced. The course is code-first and aimed at marine biologists, oceanographers, environmental consultants and fisheries or MPA managers. No programming background is needed: if you can build a pivot table in Excel, you can learn R. It starts with installing R and RStudio, setting up projects and finding help. It then teaches the fundamentals through marine examples, including vectors of CTD temperatures, data frames of survey sites, a hand-written Shannon diversity function and bleaching alert levels built with ifelse(). From there it covers importing the formats marine scientists actually meet: CSV and Excel field sheets, NetCDF satellite SST, Sea-Bird CTD files, and live ERDDAP and OBIS downloads with WoRMS name checks. The middle of the course covers the work behind every result. Wrangling uses dplyr, tidyr, lubridate and stringr, with automated QA/QC flags that keep an audit trail instead of deleting records. Visualisation uses ggplot2 for depth profiles, time series, facets and publication figures with ocean-appropriate palettes. The statistics module runs from the choice of test through t-tests, ANOVA, regression and diagnostics to Poisson and negative binomial GLMs. It ends with mixed models that treat site as a random effect so transects are not counted as independent replicates. Later modules add spatial analysis with sf, terra, marmap and leaflet, including EIA buffer screening around outfalls. They also cover time series, climatologies, marine heatwave detection with heatwaveR, Degree Heating Weeks and trend tests, and community ecology in vegan with diversity, rarefaction, Bray–Curtis dissimilarity, NMDS, PERMANOVA and SIMPER. The last core module makes the work reproducible with Quarto reports, Git, renv, parameterised site reports and FAIR data publishing. Everything runs on reef_monitor.csv, a 576-record teaching dataset covering 8 Abu Dhabi reef sites over 24 months, and ends in an end-to-end applied case on thermal stress on those reefs. The case moves from import to a mixed model, a map and a manager’s report. A technical review covers performance, Arrow and DuckDB for large data, debugging, code quality and common analytical errors in marine studies. Assessment is an online 20-question exam at a 70% pass mark, leading to a verifiable SKEDMIA certificate.