Difficult to see. Always in motion is the future – Yoda
Recent crises demonstrate that the future will not necessarily mirror the past. We face a multitude of possible, probable, and preferable futures that we need […]
At UNDP we know too well that traditional approaches to development are struggling to keep up with the dynamic speed, breadth and nature of today’s challenges. And that digital data sources are playing an increasingly important role in the way we scope, design, implement and evaluate our work.
From working with colleagues to shape and refine our thinking around data innovation and collaborations with country offices and their new Accelerator Labs, here are some of the lessons we’ve learned and what we intend to do differently going forward.
Big data from mobile phones or satellites provide an increasingly accurate picture of the movements, choices and behavior of large numbers of people, but they reveal very little about people’s intentions, drivers, experiences and emotions. In other words, we’re able to see the “what” but learn very little about the “why”. To form a complete and in-depth picture of reality, we need to combine large, quantitative datasets with much more granular, context-rich qualitative information (or “thick data”).
Pulse Lab Jakarta’s Haze Gazer, a tool that provides real-time insights on the locations of fires and haze hotspots in Indonesia, is a prime example of this approach, including big data from satellites as well as qualitative data from civic journalism videos. Similarly, to improve mobility in the city of Chisinau, UNDP Moldova complements spatial and mobile phone data with information from micro-narratives to capture the perceptions of urban commuters.
As data innovation practitioners, our belief in the sheer power of data can be naive at times. For a long time, we’ve thought that if we just provided policy-makers with sufficient data on a particular issue, they would just follow the evidence. That someone out there would use it to build apps, products or services. But if we really want to unleash the power of data, we need to go further, combining data-based methods with other innovation practices, such as systems thinking, human-centered design, collective intelligence or behavioral science that can address the shortcomings of data.
We are seeing how data is combined with nudges to drive people’s behavior, how human-centered design is used to help civil servants manage administrative data more efficiently, and how collective intelligence is used to address the shortcomings of data-based methods. For instance, UNDP Armenia has used behavioral interventions to increase the take-up rate of the cervical cancer screening programs among women and UNDP Lebanon applied design thinking to enhance local engagement of the crisis response using WhatsApp.
Many of the quick fixes we’ve fallen for in the search for solutions are based on deeper conceptual misconceptions. We tend to think in false binaries such as open vs. closed data, data supply vs demand and producers vs. consumers of data. In reality, however, people and groups might be better analyzed and engaged with based on the influence they have over a certain issue than on data supplier vs. user or similar concepts. Likewise, some data can easily be categorized as open or closed, while other cases are much less straightforward. Community groups and citizens can be data producers and aggregators themselves. Initiatives have demonstrated how data generated by citizens can be used to map pollution, support environmental justice and assist in disaster relief efforts.
We’re currently observing a lot of conceptual thinking and practical experimentation in the wider data governance community around new forms of data engagement, sharing and use, including data trusts and data collaboratives. At UNDP, we’re starting to work on new data sharing and access arrangements with partners like the European Space Agency building on experiences with projects like our Chisinau Data Collaborative.
Our data-centric view needs to put greater emphasis on context and not ignore power, politics, institutions and culture. All too often, we’re seduced by technical fixes that don’t take into account the specific culture, practices and relationships within the institutions we’re working with. In the past this has led to ‘quick wins’, but little long-term change in how data is collected, managed, shared and used.
What’s more, data itself is highly political. Who is counted matters — and who is being left out matters even more. For instance, we need to be aware that much of the data used for development work fails to take into account gender and other identity markers, which then further bakes bias and discrimination in systems. As decision-making is increasingly automated and predictive, and as human and artificial intelligence interact more, it is critical to think of the unanticipated impact of such technology. There is a lot to learn for data for development practitioners from other fields where organizations have developed and tested tools to analyze power and address power relations.
People are not mere data farms from which data is extracted to influence policy and feed the models and algorithms hidden behind firewalls. Further growth and centralization of the tech industry is increasing the disassociation between data we produce and how it is used. We need to empower the people we serve and work with — informing them about and respecting their data rights and engaging them in the collection and use of data.
There are a growing (but not yet enough) number of initiatives that involve people in data collection, use and analysis, that share ownership of data with those it’s from, and that open data to be accessible and reusable by others. One of the most widely applied methods stems from the Poverty Stoplight initiative that supports poor families in understanding and using data about their living conditions to navigate their way out of poverty.
Methods that seek to identify and build on locally-sourced solutions, such as Positive Deviance and Lead User Innovation, have been around for many years, but haven’t seen widespread use in the international development sector, mainly because of the time and labor that comes with these approaches. The growing amounts of data being generated by sensors, satellites, social media and mobile phones, open up new, more cost-effective ways for these methods to be applied.
With a positive deviance approach, for example, using large datasets to identify those outperforming their peers and their uncommon practices and behaviors could mean a new angle for data’s influence on positive development. Such an approach could be applied in a range of sectors, including infectious disease control, urban planning, deforestation and agriculture. For instance, healthcare providers could use it to understand the conditions and experiences which enable some people to be healthier than others.
The data for development space as a whole has seen its fair share of mishaps, failures and setbacks over the past ten or so years. This is partly due to overly simplistic theories of change, a naive belief in the power of data and a lack of political economy thinking. However, there has been significant learning on how data can be used more effectively to advance development.
As we reshape, refocus and refine our data innovation work, we’re working with the UNDP Accelerator Labs in using digital data to identify unlikely innovators. We’re connecting country offices with data innovation experts to disrupt our ‘ways of doing’ data for development. And will go forward, experimenting while being conscious of the limitations of data, the need to address power relations and potential unintended consequences and to put people at the heart of our data innovation work.
Interest in data and innovation? Check out our other thought pieces on the topic.
Have feedback on this article or other data and innovation ideas? Reach out to one of the contributors:
At UNDP we know too well that traditional approaches to development are struggling to keep up with the dynamic speed, breadth and nature of today’s challenges. And that digital data sources are playing an increasingly important role in the way we scope, design, implement and evaluate our work.
From working with colleagues to shape and refine our thinking around data innovation and collaborations with country offices and their new Accelerator Labs, here are some of the lessons we’ve learned and what we intend to do differently going forward.
Big data from mobile phones or satellites provide an increasingly accurate picture of the movements, choices and behavior of large numbers of people, but they reveal very little about people’s intentions, drivers, experiences and emotions. In other words, we’re able to see the “what” but learn very little about the “why”. To form a complete and in-depth picture of reality, we need to combine large, quantitative datasets with much more granular, context-rich qualitative information (or “thick data”).
Pulse Lab Jakarta’s Haze Gazer, a tool that provides real-time insights on the locations of fires and haze hotspots in Indonesia, is a prime example of this approach, including big data from satellites as well as qualitative data from civic journalism videos. Similarly, to improve mobility in the city of Chisinau, UNDP Moldova complements spatial and mobile phone data with information from micro-narratives to capture the perceptions of urban commuters.
As data innovation practitioners, our belief in the sheer power of data can be naive at times. For a long time, we’ve thought that if we just provided policy-makers with sufficient data on a particular issue, they would just follow the evidence. That someone out there would use it to build apps, products or services. But if we really want to unleash the power of data, we need to go further, combining data-based methods with other innovation practices, such as systems thinking, human-centered design, collective intelligence or behavioral science that can address the shortcomings of data.
We are seeing how data is combined with nudges to drive people’s behavior, how human-centered design is used to help civil servants manage administrative data more efficiently, and how collective intelligence is used to address the shortcomings of data-based methods. For instance, UNDP Armenia has used behavioral interventions to increase the take-up rate of the cervical cancer screening programs among women and UNDP Lebanon applied design thinking to enhance local engagement of the crisis response using WhatsApp.
Many of the quick fixes we’ve fallen for in the search for solutions are based on deeper conceptual misconceptions. We tend to think in false binaries such as open vs. closed data, data supply vs demand and producers vs. consumers of data. In reality, however, people and groups might be better analyzed and engaged with based on the influence they have over a certain issue than on data supplier vs. user or similar concepts. Likewise, some data can easily be categorized as open or closed, while other cases are much less straightforward. Community groups and citizens can be data producers and aggregators themselves. Initiatives have demonstrated how data generated by citizens can be used to map pollution, support environmental justice and assist in disaster relief efforts.
We’re currently observing a lot of conceptual thinking and practical experimentation in the wider data governance community around new forms of data engagement, sharing and use, including data trusts and data collaboratives. At UNDP, we’re starting to work on new data sharing and access arrangements with partners like the European Space Agency building on experiences with projects like our Chisinau Data Collaborative.
Our data-centric view needs to put greater emphasis on context and not ignore power, politics, institutions and culture. All too often, we’re seduced by technical fixes that don’t take into account the specific culture, practices and relationships within the institutions we’re working with. In the past this has led to ‘quick wins’, but little long-term change in how data is collected, managed, shared and used.
What’s more, data itself is highly political. Who is counted matters — and who is being left out matters even more. For instance, we need to be aware that much of the data used for development work fails to take into account gender and other identity markers, which then further bakes bias and discrimination in systems. As decision-making is increasingly automated and predictive, and as human and artificial intelligence interact more, it is critical to think of the unanticipated impact of such technology. There is a lot to learn for data for development practitioners from other fields where organizations have developed and tested tools to analyze power and address power relations.
People are not mere data farms from which data is extracted to influence policy and feed the models and algorithms hidden behind firewalls. Further growth and centralization of the tech industry is increasing the disassociation between data we produce and how it is used. We need to empower the people we serve and work with — informing them about and respecting their data rights and engaging them in the collection and use of data.
There are a growing (but not yet enough) number of initiatives that involve people in data collection, use and analysis, that share ownership of data with those it’s from, and that open data to be accessible and reusable by others. One of the most widely applied methods stems from the Poverty Stoplight initiative that supports poor families in understanding and using data about their living conditions to navigate their way out of poverty.
Methods that seek to identify and build on locally-sourced solutions, such as Positive Deviance and Lead User Innovation, have been around for many years, but haven’t seen widespread use in the international development sector, mainly because of the time and labor that comes with these approaches. The growing amounts of data being generated by sensors, satellites, social media and mobile phones, open up new, more cost-effective ways for these methods to be applied.
With a positive deviance approach, for example, using large datasets to identify those outperforming their peers and their uncommon practices and behaviors could mean a new angle for data’s influence on positive development. Such an approach could be applied in a range of sectors, including infectious disease control, urban planning, deforestation and agriculture. For instance, healthcare providers could use it to understand the conditions and experiences which enable some people to be healthier than others.
The data for development space as a whole has seen its fair share of mishaps, failures and setbacks over the past ten or so years. This is partly due to overly simplistic theories of change, a naive belief in the power of data and a lack of political economy thinking. However, there has been significant learning on how data can be used more effectively to advance development.
As we reshape, refocus and refine our data innovation work, we’re working with the UNDP Accelerator Labs in using digital data to identify unlikely innovators. We’re connecting country offices with data innovation experts to disrupt our ‘ways of doing’ data for development. And will go forward, experimenting while being conscious of the limitations of data, the need to address power relations and potential unintended consequences and to put people at the heart of our data innovation work.
Interest in data and innovation? Check out our other thought pieces on the topic.
Have feedback on this article or other data and innovation ideas? Reach out to one of the contributors:
RELATED POSTS
Recent crises demonstrate that the future will not necessarily mirror the past. We face a multitude of possible, probable, and preferable futures that we need […]
Edited for brevity by Arianna Friedman and Svetla Baeva. See full article on the UNDP Kazakhstan website. 99% of the global population is exposed to […]
Gender disparities persist globally, particularly in positions of power and higher pay, where women face cultural and professional barriers to advancement, which are often referred […]