Data Science for Nonprofits
Nonprofits can use data science in three main ways: marketing and fundraising, monitoring and implementing specific activities, and streamlining funds.
Data-Driven Fundraising and Marketing Campaigns
With regards to fundraising, big data can be used to see each prospective donor in a more comprehensive view in order to individualize interactions between nonprofits and potential donors. Because relationships with donors are among the most important factors in marketing, these insights can be used to drive decision-making in fundraising and marketing efforts, both to boost visibility of the nonprofit and to drive donor contributions to the nonprofit organization.
Insights that can be gleaned from big data and data analytics to power fundraising include the development of statistical models to optimize fundraising; identification and categorization of donors for targeted fundraising efforts; identifying new relationships; and identifying the value that nonprofits’ work has for donors.
Notably, predictive modeling can be used to identify audiences for targeted marketing. The secret to effective content marketing is to make sure that the marketing strategy is relevant to the target audience. Fundraising- and marketing- related data analytics, therefore, can play a significant role in this effort. Knowing about your target audience in terms of their preferences, purchasing and donation behaviors, relationships with others in the nonprofit world, and of course, their income level and ability to donate, can help nonprofits craft messages that are individualized and highly relevant.
Amnesty International is one example of a nonprofit that has used segmentation and predictive modelling to identify audiences for targeted marketing. As a global organization with over seven million members who work towards improved human rights, Amnesty International has recently used data analytics methods to examine patterns of behavior among donors, better customize messages to prospective donors, and determine the effectiveness of fundraising campaigns.
The Belgian branch of Amnesty International recently partnered with the data analytics platform RapidMiner to modernize reporting and analytics capabilities. Amnesty International has integrated RapidMiner software with Customer Relationship Management (CRM) in order to manage company relationships and interactions and improve business relationships. CRM is designed to improve revenue, and is typically used by companies to boost profits, but it can play a role in fundraising for nonprofits as well.
In this way, data science has allowed Amnesty International to leverage more effective fundraising and marketing campaigns in order to reach more people and reach them more effectively.
Monitoring, Implementing, Predicting, and Streamlining Nonprofit Activities
Both nonprofit and for-profit companies can use big data to boost revenue for their organization – in the nonprofit’s case, to help cover operating expenses, rather than to improve the bottom line. Another thing that sets nonprofits apart from for-profit organizations is the type of work in which nonprofits are engaged. Big data can help nonprofits improve the monitoring and implementation of their organizational efforts, just as it has helped companies improve their day-to-day operations.
Nonprofit organizations include food banks, museums, colleges, and others. Many nonprofits are engaged in a variety of efforts, on the day-to-day scale which can be benefited by big data. Predictive analytics is a major advantage of data science for nonprofits. In other words, big data can help companies accurately measure effectiveness of efforts and to tailor future efforts. For example, big data can be used by food banks to determine the need for food, and to identify donors who can meet this need in times of high demand, such as natural disasters or other devastating events. Big data can also be used by colleges to examine a student’s past performance in coursework to help figure out what students have a greater chance of graduating versus dropping out of the institution.
Many nonprofits operate via a web platform, such as the educational nonprofit Khan Academy, which means that a plethora of data is generated from users visiting the site, interacting with and downloading content, watching videos, etc. Data analytics is hugely powerful, especially in these web- and app-based contexts, to help determine what content is most popular among users, and help determine future content strategy. Data analytics can also be paired with Artificial Intelligence to help create chatbots and customer service resources that can help users find help for their most commonly asked questions, for example. There is no limit to how big data can be used in these contexts to drive an exceptional user experience.
Data analysis and visualizations from big data provide organized, comprehensive, and digestible information for real-time tracking and targeting of efforts. Data analytics can help uncover patterns and power dynamics which can be used to inform decision-making. Predictive modeling can also be used to identify potentially relevant individuals and communities who may be in need of the nonprofit’s assistance. The University of Chicago’s Data Science for Good program used predictive modeling to identify homes in the local area that likely contain lead-based paint, which could then be targeted for repair in order to reduce lead exposure, which is a known cause of developmental delays and neurocognitive deficits in youth.
Finally, big data can be used to maximize cost-effectiveness and efficient allocation of resources. For example, the Akshaya Patra foundation uses data analytics to make meal delivery to government schools more efficient.
Importantly, nonprofits can benefit both from freelance or consulting data scientists for short-term projects or rely on assistance on volunteer scientists.
If you are working in a nonprofit, you may think, “This all sounds great, but how can we put data analytics into practice?” In the next section, we review DataKind, a company dedicated to “harnessing the power of data science in the service of humanity.”
Natural Language Processing to Examine Social Media Habits
Natural Language Processing (NLP) can be used to analyze people’s social media behaviors and develop better marketing and fundraising strategies. Big data is key for decision-making in nonprofits just as it has been in the corporate world. The two main uses of big data in nonprofits are to promote effective fundraising and identify most effective and efficient use of funding. Online resources can be exploited in the digital age to power decision making via big data. Unstructured data from social media platforms can be used for fundraising, as well as characterizing and identifying audiences. Natural Language Processing (NLP) can be used to analyze text from social media postings to understand how audiences view and talk about an issue or topic. NLP can, in this way, be useful to structure fundraising campaigns. Data analytics can also be used to create targeted marketing campaigns for specific demographic groups, e.g., to identify a subgroup of people who may be most interested in the nonprofit’s cause. Sentiment analysis can be conducted using NLP to ensure that the campaign is effective among audiences, and can track social media discussions via analytics and conduct “spread the word” campaigns on social media platforms such as Facebook and Twitter.
Bringing Actionable, Data-Driven Insights to Nonprofits: DataKind
DataKind (www.datakind.com) is an organization that unites the world’s leading top data scientists with prominent social change organizations to work together on the latest trends and innovations in analytics and mathematical algorithms to “maximize social impact.” Designed to “meet organizations where they are,” DataKind programs are built upon one another. The programs are both short- and longer-term efforts, on both small and large scales. Pro bono assistance is provided for humanitarian efforts and DataKind draws on a global coalition of volunteers. DataKind’s website features volunteer opportunities, a site to submit humanitarian project ideas, and examples of completed projects.
DataKind can be seen as a case study for the use of data science for nonprofits. It acts as a volunteer hub and placement tool to coordinate the use of data science for the benefit of organizations working toward humanitarian goals. DataKind is a nonprofit geared towards working with other nonprofits and social change organizations.
DataKind has been working on many projects to help nonprofits work better. Examples include improving college success through predictive modeling for students of John Jay College of Criminal Justice in New York City, New York. DataKind built a predictive model using data from John Jay to be able to identify students who were at risk of dropping out, in order to offer timely support to help them complete their education. This project helped develop a blueprint for an application that takes student data and analyzes it to compute dropout risk for each student. In 2017, DataKind helped Washington, DC’s Child and Family Services Agency (CFSA) case workers determine clients’ eligibility for multiple programs and social services at once. The solution developed by DataKind provided clients with a portal asking a few simple questions, the answers for which could then be used to determine client eligibility for a host of programs. This solution saves CFSA caseworkers time and effort as usually, case workers must manually look through the client’s needs to suggest a case. People who need many different social services must enlist many different case workers, also, while this new solution is more centralized and can direct clients to proper resources based on their input regarding their needs.
Source: Discover Data Science