Skip to content

Latest commit

 

History

History
90 lines (60 loc) · 3.32 KB

taf_discussion.md

File metadata and controls

90 lines (60 loc) · 3.32 KB

TAF

A TAF analysis is

  • Modular - the analysis is divided into well defined steps, with a separate script for each step.

  • Structured - the scripts are run sequentially, the purpose of each script is clear, scripts are tightly linked with each other

  • Generic - each script ends by writing out files in generic formats (often *.csv and *.png) that can be used by any software to create products, such as dynamic reports and interactive websites.

  • Data provenance - the workflow describes the origin of the data, version controls the data, and documents in precise terms how the data were processed.

  • Reproducible - the analysis can be run on any computer (today or after some years) to get the same results, using snapshots of the exact software version and relative paths.

  • Standardized - TAF is widely used in hundreds of analyses in fisheries science every year, which invites collaboration and increases reviewability of the analysis.

Benefits for SOFIA

The main benefits of running the SOFIA analysis in TAF are:

  1. Analysis of this scope becomes manageable with a foundation to organize all input files, scripts, and output files.

  2. A structured modular design makes it easier to modify code without introducing errors.

  3. The structured modular design also makes it easier to read and understand the code, improving reviewability.

  4. Finally, structured model design makes it easy to compare alternative methods and to see exactly what is different from the last SOFIA analysis.

  5. Formal data management in TAF shows information about where the data came from and what is inside each data file.

  6. TAF lends itself well to parallel computing, decreasing run times.

  7. The workflow pipeline can be extended to generate dynamic report and feed into an interactive website.

  8. TAF analyses are well suited for open science, where anyone can browse and download the analysis.

  9. Reproducibility means we can run the SOFIA analysis on our laptops or online servers.

  10. Reproducibility also means we can organize close collaboration with high-impact research groups in fisheries science.

Opportunities

  • Technical infrastructure for Reproducible Workflows to organize in-house scientific analyses. The initial focus would be on SOFIA, but a similar approach could benefit other aspects of FAO work.

  • Online environment to make the analytical workflows available as open science, along with all input and output tables. This will strengthen FAO's role as the Authorative Source of Information in global fisheries and aquaculture.

  • Moreover, FAO can provide an Active Forum of Collaborative Research focusing on the state of world fisheries and aquaculture. FAO will host an online platform where data and analyses are organized, and invite selected research groups to conduct their analyses on this FAO online platform.