Nick Hamlin, a data scientist at GlobalGiving, outlines five data management trends to track in 2019 and gives you advice for how to stay ahead of rapid innovations in the field.
2018 has been a busy year in the data world, especially in the social sector. With new technologies, regulations, and questions arriving seemingly every day, it’s hard to know what’s new challenge is going to pop up next. I suspect that this constant growth and change will continue in 2019 and include some important trends. Here are five themes I’ll be looking out for in the data space next year:
This trend has been going strong for a while now, particularly as more and more of the data landscape has shifted to using open source tools. For example, at GlobalGiving, we’ve used an open source tool called Metabase to allow everyone on our global staff to ask questions of our data, build dashboards and visualizations, and share important conclusions with colleagues without needing to write a line of code or pay for an expensive proprietary software package. Similarly, companies like Amazon are making it easy for anyone to create powerful natural language processing or computer vision models without needing to build their own complex deep learning infrastructure. Tools like these make it far easier for nonprofits to include cutting-edge data-driven approaches in their work, even without a large dedicated team.
As data technology becomes more accessible, social sector data leaders will need to focus even more deliberately on effective communication about what their systems and algorithms do, how they work, what their results mean, and what unintended consequences they might have. When third-party tools obscure most of the “nuts and bolts” of how an algorithm works, it’s easy (and tempting) to simply trust that the providers of the service have taken care of everything from overfitting to inherent biases in training data effectively. Don’t let this happen! Even if a third-party service doesn’t allow for clear inspection of their inner workings (and many don’t), communicating clearly about the blind spots in a data process will help ensure that bad assumptions don’t creep into your organization’s data work. Similarly, while your tools may offer fantastic summaries and visualizations of your data, it’s up to you and your expertise in your subject are to communicate the context and narrative of your data clearly to your audience. This is one of the most challenging aspects of data science, but also one of the most rewarding, since it’s what allows people to really derive value from your data and your work.
With the formal launch of the EU’s GDPR law (along with thousands of “we’ve updated our privacy policy!” emails) in May and several high profile stories about major companies struggling with data breaches, 2018 was a big year for conversations about what privacy really looks like in the modern data landscape. Building on that trend, we’ll likely see this conversation expanding even further in 2019, as other countries explore their own versions of data protection legislation. Nonprofits won’t be exempt from these trends, so it’s worth examining what kinds of your data your organization uses to carry out its mission now so that you can continue to do so amid an evolving regulatory environment. Even in situations where it’s not necessarily legally required, organizations owe it to the communities they seek to serve to thoughtfully examine the privacy implications of their work, particularly in situations where program participants are already members of vulnerable populations.
In addition to questions about privacy, the data world is also grappling more thoroughly and publically with challenges of bias and representation. Catherine D’Ignazio and Lauren Klein’s “Data Feminism” is one of the best resources I’ve encountered in 2018 that describes these challenges and how practitioners and organizations can better navigate them. Not only are their suggestions helpful, but they’re also walking the talk, but making the draft of their work fully available online for others to provide feedback and establishing an explicit set of values and associated success metrics to ensure that the examples and voices they amplify in their work don’t reinforce structural inequality across a wide variety of focus areas. This level of introspection is difficult, but central to ensuring our data work is intersectional, equitable, and human. I hope that we’ll continue to see more and more examples of this thoughtfulness in 2019.
The pace of innovation in the data space will remain rapid, so it’s natural to feel overwhelmed by all the new possibilities and implications for your organization’s work, both good and bad. You won’t be able to integrate every new product, understand every new technology, or even define every new buzzword—and you shouldn’t try! Focus on picking out a few key ideas, tools, or themes that are particularly relevant for your work and set the rest on the back burner. The rest of your year will thank you!
Explore more 2019 hopes, predictions, and trends from philanthropy leaders across the globe.
Featured Photo: Help Kids Learn How to Code Globally by CoderDojo FoundationFind exactly what you're looking for in our Learn Library by searching for specific words or phrases related to the content you need.