Nick Hamlin lives to make social good data accessible, understandable, and actionable for everyone. As GlobalGiving’s first Data Scientist, Nick wears many hats, including leading the organization’s data strategy, building and maintaining core infrastructure, and designing experiments to evaluate program impact. He’s a highly experienced DataKinder and most recently worked on our Data Integrity project with Medic Mobile, one of our expert partners under our Frontline Health Systems Impact Practice. He’s the developer of Pando, Root Change’s platform for exploring networks in the social sector, and part of the team behind Aidsight, an app allowing anyone to easily explore international aid transparency data to unpack hidden relationships between organizations and validate data quality. Most recently, he’s also a contributor to O’Reilly’s 97 Things About Ethics Everyone in Data Science Should Know. In his past life, he worked as a reliability consulting engineer for Fortune 500 companies around the world and held National Science Foundation research positions in China and Thailand. Outside the office, he is a folk music street performer and amateur coffee nerd.
To learn more about Nick’s journey with Medic Mobile in his own words, read on!
Tell us a little bit about yourself and your background.
I came to the social sector after pivoting away from engineering early in my career, and I’ve never looked back. My undergrad degrees were in Mechanical Engineering and History and I really enjoy solving problems at the intersection of humans and technology, so getting to work on wicked problems in the nonprofit, aid, and philanthropy space has been a natural fit. Once I realized that I wanted to stay in the data field long-term, I went back and got my Masters in Information and Data Science from Berkeley’s School of Information. I did my graduate work part time while on the Operations team at GlobalGiving, and it was great to be able to immediately bring what I was learning into our day-to-day. Nowadays, I lead our data work from my seat on our Product team.
I’ve been involved with DataKind in many ways over the years, including working on the nonprofit side of a DataCorps project with GlobalGiving back in 2015. This is my first time on the delivery side of the DataCorps, and (I think) my first stint as an “official” DataKind volunteer.
Can you briefly describe the project that you’re currently working on?
Our team worked to address data quality issues in the frontline health space. Community health volunteers provide critical services and care in areas where other health infrastructure is lacking and typically rely on mobile tools to track and organize their work. These apps yield large volumes of data, but those in the Community Health space seeking to make sense of this data often find they’re not able to trust it enough to make decisions due to a variety of data quality issues.
Our team worked with Medic Mobile, the nonprofit behind the Community Health Toolkit (CHT), one of the flagship tools in the sector. Thousands of community health workers across 23 countries use the CHT, so any improvements we could make that would help others to better trust the data it produces could have substantial benefits across the globe.
Over our six month engagement with the Medic Mobile team, we conducted extensive exploratory data analysis to more precisely capture and categories the kinds of data quality issues that CHT users were encountering. These ranged from problems as simple as duplicated patient records and missing dates of birth all the way to more complex issues like miscalibrated thermometers causing specific community health volunteers to record consistently high or low temperature readings. To address these problems, we created a set of data pipelines to rigorously test for 160+ different data quality problems,allow users to explore trends in the incidence of these issues over time, and flag specific records that are out of compliance. We also put together a “cookbook” of examples and suggestions so that other organizations can more easily implement these pipelines beyond this initial engagement with Medic Mobile.
What surprised you most about the project?
“Improving data quality” can be a vague request, and it’s one that many people tend to cringe at a bit since it usually implies hazy requirements, tedious data cleansing, and problems that aren’t very interesting. This was not one of those projects. The level of creativity that the team brought to the challenge, both in terms of which aspects we focused on and the types of solutions we ultimately created, kept things engaging and moving forward. I was also surprised to discover that we didn’t need to do any of the “sales” work that usually accompanies foundational data infrastructure work, which can be a hard topic to get leadership excited about and bought into. Medic Mobile was supportive and ready to run with this project as far and as fast as we could from day one, which made it easy to maintain our momentum all the way through.
What data science skills have been most useful for this project?
This project was tightly focused on the bottom of the “data science hierarchy of needs” – solid, clean, easy-to-access data. As a result, we skipped all the data science tricks that usually make news (like cutting-edge machine learning algorithms or complex statistical approaches) and instead zeroed-in on the essentials: conducting comprehensive exploratory data analysis, engineering reliable data pipelines, and writing clean, easily extended code and documentation. Nailing these fundamentals is critical for any successful data project, but it’s even more important if we’re aiming to both solve a problem for a specific DataKind partner AND improve a core component of the overall community health sector’s data stack.
What professional skills (non-data science) have been most useful for this project?
We spent roughly three months working on this project under “normal” conditions before COVID-19 completely changed the stressors, schedules, and working environments for pretty much everyone on the planet. Every data project needs effective communication to go smoothly, but pulling one off during a generational public health crisis demands we bring a new level of mindfulness, flexibility, and grace to the way we collaborate. I’m so impressed by how every member of this team met those new demands so thoughtfully and moved smoothly through what easily could’ve been a project-derailing challenge.
What tips do you have about communicating data science findings to nonprofits most effectively?
As a data scientist at a nonprofit myself, this is something I spent a lot of time thinking about. That said, it’s a bit different sharing findings with leaders at a different organization versus inside your own: there’s less automatic shared experience and you probably don’t have as strong an intuition about what kind of stories resonate best with the decision-makers. Data quality is a challenge for everyone though, which made it easier to empathize with the frustrations that the Medic Mobile team had and frame our communications to speak directly to those problems.
What advice would you like to share with volunteers who are new to DataKind or the Data for Good movement?
Listen more than you code.
The organizations and the communities they serve have years of experience and expertise on what makes a problem tricky, what solutions might be most helpful, and what ripple effects your work could have. If you’re not actively putting the voices of those most affected by the problem at the center of your process and making sure it’s *their* process, you’ll miss out on all of that knowledge and the lasting impact that community-led work brings. The Rev. Dr. William Barber recently summarized this idea beautifully: “Change the narrative by changing the narrators”.
What’s your favorite movie?
The 1995 Tom Hanks classic: Apollo 13. This scene is my favorite:
Engineer 1: “We gotta find a way to make this [holds up cylindrical CO2 filter]… fit into the hole for this [gestures to square CO2 filter]….using nothing but that [gestures to a pile of random parts]
Engineer 2: “Better get some coffee goin’…”
Do you have a quote, song, or a piece of poetry that really inspires you? (Points if you have it memorized.)
(Come on, there are WAY too many to choose from here. These were the first two that came to mind)
The meanest dog you’ll ever meet, it ain’t the hound dog in the street.
He’ll bare some teeth and tear some skin, but, brother, that’s the worst of him.
The dog you really got to dread’s the one that barks inside your head.
It’s him whose howling drives men mad and a mind to its undoing.
-Hermes, Act II, Hadestown, Anaïs Mitchell
For as long as time endures
As long as sentient beings remain
May I, too, live
To dispel the sufferings of the world
-Shantideva
If you wrote a letter of gratitude to your future self, what would it include?
I actually do this! Well, not full-blown letters, but I definitely address comments in code or notes on JIRA tickets to “Future Nick” and will search for that to make sure I remember important details. Future Nick is usually most grateful when Past Nick has taken the time to leave detailed instructions and removed all the tiny speed bumps in a process in advance. He means well, but Past Nick also often finds himself under a time crunch (much to Future Nick’s chagrin when he discovers something missing or unclear.)
Our volunteers are the lifeblood of our mission. They’ve inspired people to use their skills in ways they never dreamed of. They’ve slayed misconceptions. They’ve shown organizations trying to make the world a more humane place how data science and AI can change the game. We’re honored (and thrilled) to feature their stories in DataKind’s Volunteer Spotlight series. Follow this series to learn about their impeccable skill sets, their work with our brilliant project partners, and what inspires them to give their time, resources, and energy to causes that matter.
