Ameen Jauhar, Data Governance Lead in CABI’s Digital Development team, examines the importance of the FAIR (Findable, Accessible, Interoperable, and Reusable) data principles. Drawing from his experience in data governance and policy, including his work in the data and AI policy sector in India, Ameen explores nuances between global and local data governance jargon and offers practical insights to improve adoption of FAIR principles in local data practices.
Why FAIR?
We live in an increasingly digitized world where the value and predominance of data mean it has been equated to ‘oil’. This means ensuring data is more accessible and robust is crucial to maximizing the value data can bring.
Here we examine how a universally recognized set of data-related principles – the FAIR data principles – apply to regional or local contexts in India and what that means for those proposing more rigorous approaches to ensuring data can be found, combined, responsibly shared, reused and accessed. Findable, Accessible, Interoperable, and Reusable, or FAIR data has become the recognized gold standard when one discusses measures to improve access and quality of datasets.
The principles of FAIR data lie within the scientific research community, which aimed to create more seamless flow and exchange of research evidence and information. Overtime the underpinning ideas of the FAIR principles have gained traction in more generic discussions around data management practices and data governance at large. In fact, recent publications have also estimated an economic cost of data that fails to be FAIR – a report published some years back pegged the value of data at approximately €10.2 billion per year in Europe – and stipulated that not ensuring interoperability and reusability of datasets made its usage most expensive and inefficient.
FAIR data is particularly important when it comes to helping to meet the UN Sustainable Development Goals of “No Poverty” and “Zero Hunger.” In line with one of the five major goals within CABI’s Medium-Term Strategy, CABI has already been active in helping partners adopt FAIR practices to improve the food security and livelihoods of smallholder communities.
For example, CABI partnered with the Australian Centre for International Agricultural Research (ACIAR) to help it develop a FAIR data strategy for its soil and land management program. This involved proposing a way forward for the data management practices for future investments planned across different work portfolios.
Furthermore, CABI’s work on the Malawi Digital Plant Health Service (MaDiPHS) Data Catalogue has been highlighted as a key resource towards greater FAIR data in the agricultural sector – particularly in gathering information on how to diagnose, prevent and manage pests, diseases and weeds which can impact upon crop yields.
FAIR versus Open Data
It is important to flag that while the underlying element of FAIR have arguably acquired a de facto universal recognition, not everyone speaks the same language. In my current role as Data Governance Lead at CABI, FAIR principles (and its jargon) are well-embedded in even my everyday work tasks. However, in my previous role working in the data and AI policy sector in India for over half a decade, I have only witnessed the use of the FAIR acronym a handful of times.
That is not to say that Indian stakeholders are oblivious to the benefits of FAIR data. The Government of India has taken considerable interest in improving these specific facets of indigenously developed datasets. In a data accessibility and use (draft) policy, the Ministry for Electronics and IT, among other things, focused on Interoperability which has been a near constant demand for India, especially with the advent of a booming AI start-up sector within the country. However, the difference in local semantics versus FAIR principles, and its impact on data governance policies and approaches, is patently evident. A useful example to demonstrate this gap is understanding ‘accessible’ versus ‘open data’.
The Open Data principles emerged around a decade ago and were aimed at enhancing transparency in governance more than anything else. However, increasingly open data has been construed as easier access to data, which has been a contentious approach to improving data access and quality. FAIR is vastly different in its definition and facets of ‘accessible data‘. It fundamentally does not subscribe to unfettered access. Instead, it aims to create adequate checks or ensure that even if a limited set of people are the intended recipients, they should be able to access such data without technical or other impediments.
FAIR’s accessible element is a more amenable standard for governments, allowing them to prescribe data sharing protocols and access control measures. This is more feasible as opposed to open data principles which (in theory) require absolute access to all data, given the sensitivity and enormity of datasets that government offices typically hold. Hence, while the Indian government may say it wants to create open datasets, a closer scrutiny of its policy and rhetoric will disclose congruence with FAIR’s accessibility principle(s) rather than the aforementioned open data principles.
Prioritizing FAIR ethos over jargon
Given the differences in local semantics and FAIR jargon, several questions emerge for consideration – what should FAIR alignment focus on? What should funders, policymakers and other bodies that aim to promote better data practices and data governance do? Should FAIR alignment be about the letter (of FAIR) or its spirit?
This conundrum may become clearer with a hypothetical scenario. Say X Intl.Ltd, an Ag-Tech developing social start up is seeking funding from a globally established philanthropic organization. The requirements of the philanthropy mandate that potential grantees must among other things, detail out their plans on how they intend to embed the FAIR principles in their project’s data asset(s), its management, and overall data governance.
X, coming from a landscape where FAIR principles are not widely recognized, is concerned about this prerequisite which may result in X’s disqualification from the funding process. However, X is familiar with issues around accessibility and reusability and aims to ensure that its data assets are designed for ease of access, and reusability.
In such an instance, will it suffice for X to demonstrate that all or most elements of FAIR are met through its plans, or must it rigidly subscribe to, and reproduce the language and format instructed by the philanthropic funder?
It is my contention that FAIR embeds certain core ideas which are articulated as the four principles. If data asset owners and managers can establish that they intend to focus on the four principles, there needs to be flexibility in how such intention is furnished and articulated. So how can this intention be demonstrated despite differences in FAIR related jargon?
DMAPs and compliance checklists
A good tool to exhibit compliance with FAIR’s spirit and intent, is a data management plan (DMAP). DMAPs can be defined as an organizational action plan covering key aspects of data management and governance. DMAPs will typically stipulate targets around data management, metadata creation, reusability, standardization of data assets being created to ensure interoperability, to name a few.
Additionally, from a data ethics perspective, DMAPs may also set action plans for how data assets will embed privacy-by-design, ensure data autonomy, fulfil collective benefits through the ethical processing of community data, and ensure the responsible use of data assets.
To create robust and actionable DMAPs, a potential organization may need to ask some questions focused on two key themes – first, what is the status quo and the existing level of FAIR maturity within the organization (both institutional and personnel level); and how does it envision to augment this maturity and culture with its setup. The latter is crucial when an organization seeks funding for data rich projects.
Going back to the example above, for X to satisfy FAIR alignment, a detailed and robust DMAP should be deemed suitable to satisfy the preconditions of the funding requirements. What would be relevant for evaluating the efficacy of such a DMAP would be an examination of its substantive intent, objectives, and action points, and not a rigid review of the superficial jargon it uses (or avoids).
Summary
The purpose of this piece is to showcase how FAIR aligned practices may not always use common language and jargon. However, determining whether such practices are meaningful (or not) must come down to substantive evaluation. This is especially important for funders, policymakers and oversight boards (within an organization) that are aiming to facilitate and promote FAIR data practices and governance. Tools like DMAPs can be practical and useful in furnishing the action and intent of an organization or grantee towards embedding FAIR principles within its data assets. As such, funders and other drivers of FAIR should advocate for the development and adoption of such tools, to gauge substantive intent, while ensuring flexibility around FAIR jargon. This will prevent the FAIR alignment process from becoming needlessly pedantic and cumbersome.
Additional information
Project page
Find out more about CABI’s work on enabling FAIR and responsible data practices through the project page ‘Enabling FAIR data sharing and responsible data use’
FAIR Journal
Learn more on the topic through the FAIR Journal – a bite-sized digital magazine-style resource of key insights, emerging knowledge, advocacy, and learning around the FAIR Principles curated by CABI.
Relevant stories
CABI partners with ACIAR to develop FAIR data strategy for its soil and land management program
CABI shares expertise in FAIR data at SciDataCon conference as part of International Data Week
Story of impact
Data access project prompts creation of new data-sharing law in Ethiopia
Main Image
Credit: da-kuq/iStock
Related News & Blogs
IRRI and CABI collaborate on biocontrol workshop for sustainable agriculture in Odisha
This article was originally published by the International Rice Research Institute (IRRI) To promote the use of biocontrol agents and biopesticides and mainstream Integrated Pest Management in the rice-based ecosystem in Odisha, the International Rice…
24 October 2024
