How to draw a data flow diagram

[vildead]

Recipe Type

Hands-on, Technical Guidance

Audience

Data Engineer, Data Manager, Data Scientist, Data Steward

Executable Code

No

Main Objectives

Give users clear guidance on how to design and make data flow diagram

This recipe should extend the related RDMkit page(explaining why) by practical steps (explaining how)

Graphical Overview

No response

Requirements

Hands-on experience with visualization of data flow
Basic knowledge of graphical tools

FAIRification Objectives, Inputs and Outputs

Inputs: In general, the process starts with project description/proposal but other inputs can be included.
Outputs: Data flow diagram you can share with partners

Table of Data Standards

maybe UML but it tends to be overkill for research projects

Introduction

see the RDMKIt page on that: https://rdmkit.elixir-europe.org/creating_dataflow_diagram#how-to-create-a-data-flow-diagram

Ingredients

any graphical tool allowing user to work with vectors

Recipe body or step-by-step guidance

Still work-in-progress:

• List all assets.
• List all partners and actors​.
• List all processes​.
• Define data sources​ - such as instrument, patient, collaborator, lab.
• Define final data locations (sinks)​ - for example, archives, repositories, external processes, ingestion zones.
• Start drawing visual elements representing the outer interface (sources and sinks) and move inwards.
• Iteratively refine the diagram based on items which were not yet included.

Conclusion

TODO

What to read next?

references to visualization in research? Martin Krzywinski series on best practices?

References

tbd

Authors

tbd

[marikapop]

I would add specific instructions to Ingredient section, like in RDMkit:
Open source: draw.io, Inkscape, Libre Office Draw.
Licensed: MS Visio, Miro.com, Corel Draw, MS PowerPoint, Affinity Designer or Adobe Illustrator.

[vildead]

I think it might be even moved from RDMKit to the recipe.

RDMkit page then becomes the context and in the solution, we can list advices on what to do with such a diagram or how to make the most use of it (include in DMP, disseminate to partners, ...)