Harnessing the Power of Data Science for Social Good
It’s hard to ignore the role of data in the current pandemic: tracking the spread of the coronavirus, testing people around the world for COVID-19 infection, tracing contacts to determine exposure, and developing the new vaccines intended to combat the virus. From the Johns Hopkins University COVID-19 dashboard created in the early days of the pandemic to vaccine efficacy, approval, and availability charts, it’s fair to say that more data is being analyzed and consumed than ever before.
While a lot has been written about why and how for-profit corporations use data science to more accurately target customers and make better — yes, data-driven — decisions overall, it might surprise you to learn how many not-for-profit organizations rely on data science in order to propel their missions forward.
Here are a couple of impressive examples of data-driven non-profit efforts:
The Trevor Project, a non-profit organization focused on ending suicide among LGBTQ youth, has experienced a significant increase in demand for its crisis services since the onset of the pandemic, at times more than double its volume earlier in 2020. The Trevor Project’s confidential crisis services, which began with a single telephone lifeline in 1998, now include SMS and web browser-based chats — all 24/7 and for free.
With more LGBTQ youth in crisis reaching out than ever before, the organization looked to technology to assess suicide risk level for each of them. Using Natural Language Processing in the crisis contact intake process, youth at highest risk are now connected to trained crisis counselors faster, enabling the organization to have an even greater impact on the youth they serve. By approaching the task as a binary text classification problem, the team at The Trevor Project used responses from the intake form as the input for the model:
● Have you attempted suicide before? Yes / No
● Do you have thoughts of suicide? Yes / No
● How upset are you? [multiple choice]
● What’s going on? [free text input]
The combination of answers are important in assessing the suicide risk of LGBTQ youth — and data science plays a key role. The Trevor Project uses data culled (and scrubbed of PII) from TrevorSpace, its safe-space social networking site for LGBTQ youth to find peer connections, in order to train the AI model. The output is a binary classification: whether to place the youth in the standard queue or the priority queue. As counselors become available, they can provide support to LGBTQ youth who are at the highest risk for suicide faster than ever before. In this way, data science is helping to save lives, a key component of The Trevor Project’s mission.
Another interesting use case of data science can be found at DoSomething.org, the largest not-for-profit organization exclusively for young people and social change. With millions of members around the world, DoSomething.org’s digital platform powers real-world impact, as members join the organization’s volunteer, social change, and civic action campaigns.
So how does data science figure into the DoSomething.org mission? To help the organization understand which campaigns are most popular, why members are attracted to specific campaigns, and ultimately, how DoSomething.org can affect long-term retention outcomes.
The organization ingests data from multiple sources into its data warehouse for analysis, both internal data as well as third-party data, using automated data integration. For example, DoSomething.org brings in web event data from Snowplow into its Postgres data warehouse, so the data science team can look at campaign event data alongside product feature events and marketing actions to understand what contributed to campaign success.
With a team of only three individuals, including one data engineer and one data scientist, it’s important for others at DoSomething.org to self-service their specific analytic use cases, so the team uses Looker to expose data sets to their internal stakeholders.
In perhaps the most important demonstration of its use of data science, last year DoSomething.org registered over 250,000 young people to vote ahead of the 2020 Presidential Election by using data to ensure the organization was putting the right resources toward the efforts that were driving those registrations, and to optimize the campaign to reach that goal.
Given the growth of data science in general, it’s likely that we’ll see many more examples of data science being used to affect social change and support efforts for marginalized groups. If these two use cases are any indication, we’re in for a wave of organizations turning to data science to further their missions.
