CIAT’s data and information management team strives towards becoming the knowledge hub for scientists. The group also aims for CIAT staff to embrace an open science culture and becoming 100% compliant with Open Access goals without compromising quality of work.
In July 2015 a group of 30 data managers from CIAT HQ gathered for the first time to understanding who is doing what, get a sense of the bottlenecks and opportunities, and eventually creating a community of practice to boost progress in data management and open access.
We started by hearing different perspectives on data management. Luis Augusto Becerra who is a cassava geneticists and senior scientist and who has been very proactive in data management, thinks that data can enable change because it generates new knowledge, supports decision making for future investments, and is a way of integrating teams who take ownership of research through collective data management practices. He also alerted that any data management tool or solution needs to provide a win-win situation so that additional work load is justified. Sharon Gourdji, who leads the crop modelling team emphasizes the need for them to have access to a great volume of data so as to forecast agroclimates. Volume of data is given a higher priority to quality of data. The team also prefers to re-use existing data rather than recollect data. David Abreu is CCAFS knowledge and information manager. He framed data management in the context of research project management, and the monitoring and evaluation of the achievement of research outputs. A comprehensive platform is being developed for that purpose.
The second part of the day was dedicated to map current data management practices across all programs. We asked participants to identify by area or unit, their data management functions, current tools for data acquisition, management, storage, sharing, and analysis. The mapping exercise reveals, without much surprise, very different approaches for biophysical, socio-economic, geo-spatial and climate data, without that much difference in the tools in use for data analysis.
When we asked about the top three bottlenecks for smart data management it appeared that many have real trouble with assuring shareability of data through consistent metadata, this is because of lack of time, clear guidelines and procedures. Data managers also struggle with data structuring and are worried about storage capacities.
Next steps are to share the mapping exercise for corrections and further input from the regions. The group will then conclude on a priority collection exercise, which may lead to the establishment of connections between the many data sources that are available, but not yet fully accessible and applicable.