To address today’s most pressing development challenges, we must make better decisions. That is why quality, timely data are so important. Without quality data, evidence-based decision making is a lot harder and must rely on intuition expertise or theory.
While the potential benefit from new and diverse data sources has been talked about for some time, crossing the gap between expectation and reality remains a work in progress. As the quantity and diversity of data collected and shared continues to increase rapidly, many new opportunities and challenges are emerging. In addition to traditional sources of data, such as surveys, there is satellite imagery and mobile money transfers. By now, these opportunities are becoming familiar. At the same time there is growing concern around emerging risks in the use of data such as erosion of individual privacy, bias and misuse.
The rising tide of data also presents a huge challenge. Recognizing and addressing these issues is essential in order to cross the reality-expectation divide.
Since the launch of the first Human Development Index (HDI) in 1990, UNDP has been at the forefront of the use of data in guiding the development agenda. Our colleagues working across poverty, governance, climate, crisis recovery and more all rely on data to make decisions every day, so it is clear that UNDP needs to continually evolve and improve how we work with data. This is why we are in the process of developing a comprehensive Data Strategy, to be released in early 2021, and underpinned by a set of 8 Data Principles.
These 8 Data Principles, outlined below, seek to articulate the values we hold when harnessing data both new and existing sources.
- Safeguard personal data; Embed 'privacy by design' into all data practices. Get informed consent and ensure data are anonymized before publishing.
- Uphold the highest ethical standards; Anchor data practices in the UN Charter and international human rights frameworks. Ensure that data processes and partnerships serve the public good.
- Manage responsibly; Practice effective stewardship and governance to ensure sound data quality, security and accountability. In accordance with relevant institutional policies and regulations.
- Open by default; Make data available as widely as possible. No ‘data hoarding’.
- Plan for reusability and interoperability; Maximize the value of our data by ensuring it is usable in multiple domains. Make use of open standards and machine-readable formats in order to improve interoperability.
- Empower people; Provide people with the technology and data literacy skills to be able to effectively work with and understand data. Support governments, local communities and civil society partners to strengthen data and statistical capacities.
- Expand frontiers; Explore emerging practices and innovative technologies to increase data availability and expand coverage of under-represented groups through data collection and disaggregation.
- Be aware of limitations; Assess gaps, risks and bias in the use of data. Actively question blind spots and potential negative implications of data use.
The publication of these Data Principles this month is just the first step.
The greater challenge is to enable people to enact these values every day. The Data Strategy will help us put the principles into practice by providing the tools, processes and frameworks to implement the principles when working with data. This will enable us to take advantage of the opportunities that data provides to improve our work, while at the same time ensuring we are one step ahead of emerging risks.
So, what does the future of data-driven development look like? It looks like a world where data silos are broken down, where persistent data enables feedback loops to continually improve outcomes, and where individual rights are safeguarded. Getting there requires that people across UNDP are empowered to work with data; that they have the skills, tools and resources necessary to work with data effectively and responsibly. As we look to 2021, this is what we must strive towards.