Man plans, and God laughs. The last two years of a global pandemic have brought this Yiddish proverb to life as so many unexpected events have occurred—and not occurred—as schools and offices and entire economies shut down. Above all, we mourn the unforeseen loss of more than 5 million people worldwide to COVID-19. This loss, and the associated legacy of rising poverty, has heightened awareness of the individual and societal costs of health inequality as sharp disparities are highlighted both within industrialized nations and around the world. The enduring pandemic has also brought changes in social and work norms as organizations reshape and retrain themselves for new behaviors, like a rise in remote work. In our own field of data science for social impact, here are just a few of the shifts our team has observed in 2021: 

• Increased awareness of what data is available—and the high cost of the large swaths and biases inherent to the data that is missing  
• A sense of urgency in incorporating data in problem-solving efforts to tackle broad, systemic challenges like those brought and exacerbated by the pandemic  
• Renewed interest in cross-boundary collaboration to avoid duplication of data-focused efforts when solving critical, time-sensitive problems  
• Broad interest in the development of purpose-driven data professionals to meet the growing need for analysis of health and related data

What else might emerge in data science for social impact in the coming year? We’ve asked a few of our colleagues in the field to share their predictions and possibilities for 2022.  

George Kibala Bauer, Senior Advocacy and Insights Manager, GSMA, focused on the need for more and better data fueled by data-sharing partnerships.  
 
“Data scarcity makes it difficult for cities to adequately respond to critical challenges such as rapid urbanization, climate change, and inequality. To access more timely, accurate, and cost-effective insights from the private sector, cities will be at the forefront of public-private data-sharing partnerships and continue to pioneer new use cases for innovative data sources such as mobile big data. It will be important to overcome barriers to insights from data partnerships leading to tangible policy impact, and help these partnerships scale beyond pilots.”  

Sonja Kelly, Director of Research and Advocacy, Women’s World Banking (a 2021 data.org Challenge awardee), and Jake Porway, Fellow at data.org both pointed to a greater role for regulation in how data science is applied.   

“In 2022, we can expect to see increased regulatory attention on data science applied to those industries highly vulnerable to bias, and this isn’t a bad thing. Increased proportionate regulation that is gender-sensitive will level the playing field and make ethical use of data a consumer right rather than a privilege, thus helping to advance algorithm fairness.”
– Sonja Kelly 

“2022 will be the year AI ethics turns up or tunes out. The tech headlines are dominated by conversations about AI reinforcing systemic inequalities and of companies hiring AI ethicists, but the results have been mixed. This year the field will have to make some major strides, either through regulation like the Filter Bubble Transparency Act, or industry standards. Otherwise, people will become disillusioned that it’s just another version of corporate brandwashing, with little to show for it.” 
– Jake Porway 

Olasimbo Sojinrin, Nigeria Country Director, Solar Sister (another Challenge awardee) focused on increased work incorporating data in last-mile solutions.   

“Today, Solar Sister uses data at all levels by demystifying it and presenting it in a relatable and digestible manner, making it very practical to use by women in rural communities. We have found this focus on clarity opens new opportunities for greater impact in advancing energy transition for last-mile communities. We see this last-mile focus as an important component of success, and as an enduring trend that will continue in 2022 and beyond.”

Bubacarr Bah, German Research Chair, Mathematics and Data Science, African Institute for Mathematical Sciences (AIMS), pointed to the cross-boundary nature of social impact work, and the technology we will need to develop and maintain to support that collaboration.  
 
“Increased collaboration of social impact organizations globally will necessitate the development of better federated learning algorithms.”

For myself, I hope that as a sector we double down on simplicity and clarity of communication—both in framing the problems to be solved, and in getting specific about the role of data science as a solution. To do the hard work required for tackling the global problems illuminated and worsened by the pandemic, we need to engage people and partners across sectors with varying degrees of familiarity with global development and data science. It’s too easy to default to jargon and acronyms; instead, let’s try to convey the data opportunity for the social sector as simply and clearly as possible, and bring everyone along on the journey to impact.