An increasingly popular way to make science accessible, reproducible, scalable, and replicable is through Open Science and the development of open-access tools (Lowndes et al., 2017).
The most popular open-access platform for analysis and reporting of data is R (R Development Core Team, 2016). R provides more than 5000 packages for data exploration, visualisation, modelling and statistical analysis. In addition, the R Shiny package allows for the creation of web tools, allowing users to interact with science outputs and methods, without any requirement for software (only web access is required) or a need to understand complex code. Another advantage of R Shiny is that it allows scientists to directly create web tools without any of the disadvantages associated with using 3rd party programmers (e.g. high cost, difficulty communicating ideas or interpreting science).
Within Cefas (Lowestoft and Weymouth), a recent survey suggests at least 12 such tools have been created. These include the Marine Protected Areas Management tool, and the Marine Aggregates App (https://benthosapps.net/ma_tool/), a collection of tools used by the aggregates industry for compliance monitoring. These two examples powerfully demonstrate how stakeholders (industry/regulator/SNCBs) and the public can interact with the Cefas science both for data access/visualisation and innovative analyses/reporting.
Recognising the need to be able to host such tools, Cefas science computing has recently purchased a Shiny R server. To ensure Cefas tools/applications are developed and deployed in a way which satisfies Open Science criteria, we are proposing a framework (Fig. 1). This framework will ensure that tools are based on solid foundations of: i) access to data using Cefas systems, ii) high quality, open access peer review publication, and transparency of methods using code repositories such as gitHub. Developed tools can be brigaded into applications to meet needs of specific groups (e.g. windfarm industry). These applications will ensure Cefas has a leading role in management and science development for particular activities.
The framework workflow is designed to produce four different science outputs during its implementation: i) databases, ii) scientific papers or reports, iii) methods and iv)web applications.
- Databases servers: The data is organized and storage in Relation Database Model Server , therefore the data used for research would be accessible by R analysis software as well by the web applications .
- Scientific publications: The result of the analysis using scientific methods are publishes as a scietifc report or paper. Often these publications require attached supplementary material with the scripts used for the publication. There are Open Access journals
- Methods/modules: The scripts created during the study and analysis of the data can be organized as a function within R packages. This packages include different methods used during the research process , make them public would support transparency and re-usabilty of them. These scripts could be tested , improved or re-used by open science community .
- Web applications: A clilection of methods used for a specific final users need can be accessible by a web user interface. Web applications avoid the problem of installation and of data loss , since the app is not host in local computers and installed and maintained in the remote server computer .
The framework describe the web applications as a collection of multiple methods, the methods as ordered scripts available in R package format used for paper/report writing , the scientific papers as a result of data analysis and the database server as the organized collection of these data used for research.
The current computer , R and R Shiny details are provided to support application developers. Applications or tools developed with the intention to use the Shiny
Server have to compatible with the following system specifications:
|Redhat Linux GNU
||R version 3.5.1 (2018-07-02)
- Data management:
The data underlying the web applications on this Open Science Portal source data via links
from the Cefas Data Hub , or from external database sources (Terms and Conditions may apply
unless those listed under Open Government Licence). Further spatial and non-spatial data
needed to run the applications is also made available from Open Source PostgreSQL Database
Servers hosted at Cefas. For more information please refer to the Cefas Data Management
Policy HERE, or contact via .
For more information please refer to the Cefas Data Management
Policy here or contact via email@example.com.
We strongly recomend the use of Open Source PostgreSQL/PostGIS . It is a open source database suitable for spatial data management and analysis PostgresSQL site
- R Library: In regard to the utilisation of R libraries , they need to be available on CRAN repository for the R version given in service description.
It will be needed a list of the dependencies libraries of your application , please send with your application description.
Security Assurance: The applications have to go through a proccess of information and security assurance when they are planned to
be publicly available. It will depende on the type of content and services taht the application provides.
To understand more about htis process or go through it , please contact Richard Page
This list of requirements may change based in the experience we are gaining with the implementation process.
Projects interested to build a R Shiny application or integrate an existing one into the Cefas Openscience framework, please contact firstname.lastname@example.org .
Further explanation about the openscience framework will be provided soon .
The linux server VM where R Shine Server is hosted have been provided by Cefas Science Office and Cefas IT program. Contact email@example.com
The installation and configuration of the Shiny Server community version and the NGINX server have been done by Roi Martinez.
Each web application have their own development and maintenance responsable. To report any issue related to the applications please refere to the contact person provided with in the applicaiton description