About this site
On this page
This page explains how Reef Code is built and where its data come from. It describes the continuous-integration workflow that re-runs every analysis on each render, names the two GitHub repositories behind the site (reefcode for the code and analyses, and a private repository of internal data copies), and points to the plots, captions, and download buttons on the analysis pages. It carries no monitoring-program data, no figures, and no downloadable products of its own.
We built this website in Quarto and R. Every time we render the website (turning code into an HTML webpage), R re-runs every line of the analysis. This approach, known as continuous integration1, keeps the project living: we add and modify scripts, new variables, and new time points, and the results update on the next render. When a change breaks something, git and github (version control) let us restore the entire project to the last working state.
Behind the scenes are two GitHub repositories:
-
reefcode: https://github.com/laurenkolinger/reefcode-
reefcoderuns all data processing and renders the website you are viewing. It holds the scripts that pull from the internal data repository (below), merge the sources into independent variables (drivers) and dependent variables (resilience metrics), and then aggregate and analyze those variables to determine the main drivers of resilience in the region.
-
-
VIRRSrawdata: https://github.com/laurenkolinger/VIRRSrawdata- A private repository of internal copies of data from various sources (individuals, institutions, agencies, and organizations). Reef Code does not redistribute raw data: these internal copies are working inputs only, and the original, publicly obtainable source for every dataset is listed on the Data Sources page. Each file carries a metadata sidecar (
.txt) recording its original filename and its source.
- A private repository of internal copies of data from various sources (individuals, institutions, agencies, and organizations). Reef Code does not redistribute raw data: these internal copies are working inputs only, and the original, publicly obtainable source for every dataset is listed on the Data Sources page. Each file carries a metadata sidecar (
This approach merges many datasets from different sources into one analysis, transparently and reproducibly, with no confusion about where the data come from or what they mean. It also supports easy collaboration and continued updates as monitoring programs add new years.
The study
Reef Code grows out of a long-term study of coral reef resilience in the U.S. Virgin Islands. Resilience is the capacity of a reef to withstand disturbance (resistance) and to recover afterward (recovery). Caribbean reefs recover slowly, and they face warming seas, coral disease, and pressure from fishing and land-based pollution.
The study draws on three monitoring programs (TCRMP, the U.S. National Park Service, and California State University Northridge) across 42 to 50 reef sites spanning 3 to 63 meters depth. It tracks coral cover and diversity, alongside the associated fish and invertebrate communities, then tests which drivers explain the most variation in resilience. The drivers fall into two groups: external conditions such as temperature, depth, currents, waves, and management, and internal conditions such as coral health, coral interactions, herbivory, and macroalgae. The analyses anchor on the 2005 thermal stress event and link drivers to resilience through a modeling approach adapted from Pacific reef studies.
The analysis pages linked from the top menu carry the data, figures, and results behind this overview.
No coding experience required
We wrote this website to stay accessible to readers who do not code. Each page carries enough formatted text to follow the analysis without reading the code.
Every page follows the same layout. The top of each page gives a short summary of what that section does. The bottom of most pages shows plots of the generated data, each with a caption, alongside buttons to download the data and explore it yourself. We fold most code away by default to keep the pages easy to read.
This project is a work in progress, so please ask if you have any questions. The long-term reef monitoring behind these pages runs on years of fieldwork by the team.
version 1.0.0 • published