The 2030 Agenda for Sustainable Development, adopted by all United Nations Member States in 2015, constitutes a blueprint of the numerous challenges that society is facing today – from ending poverty, hunger, and gender inequalities, to combating climate change and building sustainable cities. To reach the 17 Sustainable Development Goals (SDGs) by 2030, governments and international organizations need to be able to monitor progress, estimate the impact of potential interventions, and make forecasts on how the situation is likely to evolve. Today, scientific communities, NGOs and international development agencies are extensively exploring how Big Data, Machine Learning (ML), and Artificial Intelligence (AI) tools can be used to address these challenges. For instance, satellite imagery, mobile phone, and social media data have been used in combination with different computational techniques to estimate poverty, to predict population displacements in the aftermaths of a natural disaster, or to quantify the impact of human mobility during an epidemic outbreak. However, several challenges and limitations have been encountered, from data representativeness and bias, to models not being robust, explainable, or transparent.
We are looking for contributions that question the current approaches and study whether data-driven technologies are ready to be used for sustainable global development. The goal of this Research Topic is to showcase the current state-of-the-art in understanding and addressing the limitations and challenges encountered by scientific communities, international organizations, and NGOs, while investigating the use of Big Data and Machine Learning methods to tackle humanitarian and development issues such as mapping and predicting poverty, food insecurity, epidemic spreading, migrations, conflict and any kind of inequalities, from gender to health. This also includes identifying, describing, and mitigating issues relating to who develops the various data-driven technologies, for what use cases, how data is extracted, to how local stakeholders and data science ecosystems are engaged.
Submission can address, but are not limited to, the following topics:
- Biases and representativity issues in Big Data for development and humanitarian work
- Limitations and challenges in using mobile phones, social media, satellite imagery, or any other digital data as a proxy for human activity and mobility (including mechanisms of data generation of how socioeconomic, cultural, and ethnical inequalities shape big data)
- Limits of transferability of ML & AI models to different geographical and/or socioeconomic settings
- Limits of validity of ML & AI models over time
- Challenges in models’ explainability, interpretability, and transparency
- Challenges in transitioning research efforts into operational near-real-time tools
- Issues of using single data-driven metrics to quantify model accuracy and sustainable development
- Issues relating to inclusivity of solutions (where are methods developed, by whom, for whom, and were local stakeholders engaged)
The 2030 Agenda for Sustainable Development, adopted by all United Nations Member States in 2015, constitutes a blueprint of the numerous challenges that society is facing today – from ending poverty, hunger, and gender inequalities, to combating climate change and building sustainable cities. To reach the 17 Sustainable Development Goals (SDGs) by 2030, governments and international organizations need to be able to monitor progress, estimate the impact of potential interventions, and make forecasts on how the situation is likely to evolve. Today, scientific communities, NGOs and international development agencies are extensively exploring how Big Data, Machine Learning (ML), and Artificial Intelligence (AI) tools can be used to address these challenges. For instance, satellite imagery, mobile phone, and social media data have been used in combination with different computational techniques to estimate poverty, to predict population displacements in the aftermaths of a natural disaster, or to quantify the impact of human mobility during an epidemic outbreak. However, several challenges and limitations have been encountered, from data representativeness and bias, to models not being robust, explainable, or transparent.
We are looking for contributions that question the current approaches and study whether data-driven technologies are ready to be used for sustainable global development. The goal of this Research Topic is to showcase the current state-of-the-art in understanding and addressing the limitations and challenges encountered by scientific communities, international organizations, and NGOs, while investigating the use of Big Data and Machine Learning methods to tackle humanitarian and development issues such as mapping and predicting poverty, food insecurity, epidemic spreading, migrations, conflict and any kind of inequalities, from gender to health. This also includes identifying, describing, and mitigating issues relating to who develops the various data-driven technologies, for what use cases, how data is extracted, to how local stakeholders and data science ecosystems are engaged.
Submission can address, but are not limited to, the following topics:
- Biases and representativity issues in Big Data for development and humanitarian work
- Limitations and challenges in using mobile phones, social media, satellite imagery, or any other digital data as a proxy for human activity and mobility (including mechanisms of data generation of how socioeconomic, cultural, and ethnical inequalities shape big data)
- Limits of transferability of ML & AI models to different geographical and/or socioeconomic settings
- Limits of validity of ML & AI models over time
- Challenges in models’ explainability, interpretability, and transparency
- Challenges in transitioning research efforts into operational near-real-time tools
- Issues of using single data-driven metrics to quantify model accuracy and sustainable development
- Issues relating to inclusivity of solutions (where are methods developed, by whom, for whom, and were local stakeholders engaged)
