These are two ideas that shape everything the DCDC Network does. Neither requires a technical background to understand, and both have practical consequences for anyone who works with data, whether in a university, a government office, a hospital, or a conservation project.
FAIR is an acronym for four qualities that data should have if it is going to be genuinely useful, not just to the person who collected it, but to others who might need it in the future. The principles were published in 2016 by a group of scientists and data experts who recognized that too much research data was being collected, used once, and then effectively lost.
FAIR does not mean data must be public. Data can be FAIR and still have restricted access for good reasons (patient confidentiality, for example). What FAIR means is that the data is organized, described, and stored in a way that makes it possible for others to find it, understand what it is, and use it if they have permission.
Data is findable when it has been given a unique, persistent identifier (like a DOI) and is described with enough detail that a search can locate it. Think of the difference between a book sitting in a pile on someone's desk and a book catalogued in a library system. The information might be identical, but one is findable and the other is not.
A researcher at the University of Curaçao completes a study on coastal water quality. Without findability, a colleague in Aruba working on a similar question has no way of knowing this data exists. They start from scratch, duplicating effort and cost.
Accessibility means the data can be retrieved through a standard, open protocol. It does not necessarily mean anyone can download it freely. It means there is a clear, documented process for requesting access, and the metadata (the description of what the data contains) is always available, even if the data itself is restricted.
A government ministry collected economic data five years ago, but the analyst who managed it has since left. The files are on an old hard drive, in a format that requires software the ministry no longer has. The data exists but is no longer accessible. Proper accessibility practices would have prevented this.
Data is interoperable when it uses standardized formats and vocabularies, so that datasets from different sources can be meaningfully combined. If one hospital records blood pressure in millimeters of mercury and another uses a proprietary coding system, combining their data becomes a manual translation exercise rather than a straightforward merge.
Three islands collect tourism statistics, but each uses different category definitions, different time periods, and different file formats. A regional analysis requires weeks of manual harmonization before any actual analysis can begin. Interoperability standards would make cross-island comparison routine rather than heroic.
Reusability is where the other three principles come together. Data is reusable when it is well-documented, clearly licensed (so people know what they are permitted to do with it), and meets community standards for quality and context. Without reusability, even accessible data can be useless because nobody knows what the columns mean, how it was collected, or whether they are allowed to build on it.
A research team publishes a dataset from a food security survey, but the file contains only numeric codes with no legend, no description of the sampling method, and no license. A policy analyst who could use this to improve food distribution planning cannot, because without context the numbers are meaningless.
Reproducibility is a separate concept from FAIR, though the two reinforce each other. Where FAIR is about making data usable, reproducibility is about making the analysis transparent. It means that if you hand someone your data and your analysis steps, they can follow those steps and arrive at the same results. That is it.
This matters because research findings influence real decisions: where to allocate public spending, how to design health interventions, what policies to implement. If the analysis behind those decisions cannot be verified or repeated, we are asking people to trust conclusions on faith rather than evidence.
Reproducibility is not about distrust. It is about transparency. A reproducible analysis is one that can be checked, improved, extended, or corrected. An irreproducible analysis is a dead end.
A researcher runs an analysis by clicking through menus in a statistical program. They get a result and report it. Six months later, a reviewer asks how a particular variable was handled. The researcher cannot remember exactly which options they selected. The analysis cannot be verified or repeated.
A researcher writes a script that imports the data, cleans it, runs the analysis, and produces the output. Every step is documented. Six months later, anyone can open the script, see exactly what was done, and run it again. The analysis is transparent and verifiable.
This is one reason the DCDC Network's training program emphasizes tools like R over point-and-click software. Script-based analysis creates a complete, readable record of every decision made during the analysis. It is not the only path to reproducibility, but it is one of the most practical ones.
Small islands face a particular version of these problems. Institutional memory is fragile when staff turnover is high and teams are small. Data collected by one administration may become inaccessible to the next. Research conducted with external funding often leaves with the external researchers when the project ends.
When budgets are limited, duplicating data collection that has already been done elsewhere on the islands is an expense the region cannot afford. And when a single dataset might be the only source of information on a given topic for a given island, making sure that data is findable, accessible, interoperable, and reusable is not an abstract principle. It is a matter of institutional resilience.
The DCDC Network exists to help build the practices, skills, and infrastructure that keep knowledge in the region and make it available to the people who need it.
FAIR principles and reproducibility are not just concepts we talk about. They shape the network's training program, the tools we teach, and the way we work together. Our courses use open-source software so that every participant can continue working with the same tools after the course ends, at no cost, from any island. The materials we develop are openly available for anyone in the network to reuse and adapt.