Non-Profit Data Science Examples


We are all hearing about the importance of our data. In fact, when asked 59% of non-profit organisations wanted to make more effective use of their data. We all want to be taking the journey towards better data health throughout our organisations, but can feel that their are obstacles in our way at the same time. At Analytics in Action Ltd we often hear feedback from our clients that generating ideas as to how data science can be used in support of the organisation can be difficult. In response we are going to use this article to go through some useful examples of organisations who have successfully used data and analysis to make improvements to how they work. Our hope is that it can provide inspiration for how you can make better use of data within your own non-profit and perhaps be used in support of idea generation of your own. 

Legacy Donation Modelling, Cancer Research UK

In 2018, Cancer Research UK used statistical modelling to help them better understand the legacy market and the main drivers of legacy income (source).

AiA Thoughts: 

This is a great example of using data, which is applicable to any organisation where legacy donations make up a significant portion of their total income. In the UK nationally, legacy income makes up just over 6% of total income (source). For some organisations, however, it is not uncommon for it to make up a quarter or even a third of total income. In organisations where legacy donations do make a significant proportion of total income, when predictions or forecasts go wrong it can impact operations and even cause issues such as redundancies or closing of operations. In the past year we have seen instances of charities citing an unexpected drop in legacy donations as one of the main reasons for why they were proposing closures. Better predictions of legacy income can help plan for the future and support us in identifying potential shortfalls early so that you can make alternative arrangements if necessary. 

Doubling social impact, Educate Girls and IDinsight

Educate Girls is a non-profit which is working to improve enrollment and learning outcomes for girls in primary education in India. Their existing programmes had shown very positive outcomes and they were looking to expand into other villages. With limited resources, they were unable to expand into all areas at once. Using machine learning, they were able to identify which villages could allow them to target the most out of school girls as possible and consequently maximise their impact (Source).

AiA Thoughts:

This is a promising and exciting use of machine learning within the sector. Prioritising how to use limited resources is something we all need to take into account at some point. One of the main challenges of becoming data conscious is the cost incurred. This investment is a requirement, yet in the long run it is extremely effective in identifying efficiencies, reducing our costs and helping to maximise the impact made with the resources available. Some important questions we can continue to ask ourselves are: 

  • Do we know which of our services are delivering the most impact?
  • Are we regularly reviewing our data to see where we can make cost saving efficiencies without doing harm to our work?
  • Do we have the data we need so we can make informed decisions about the future; including what new projects we should take on that will make the most impact? 

When we can comfortably answer these questions we can be reassured that we are in better data health throughout our organisation and better informed in our decision making capabilities.

Prioritising Messages, Crisis Textline

Crisis Textline provides 24/7 support for people in crisis by text message. The organisation is always striving to use the data it generates more effectively to deliver its mission. See the founder talk about it more here.

Crisis Textline wanted to create a system whereby those people texting in who were at imminent risk were put in touch with a crisis counsellor as a matter of priority, rather than putting them in a traditional queue. Using the contents of a person’s initial text message along with post conversation surveys completed by the counseller, they are able to identify the words and phrases that those at imminent risk use and ensure they are contacted by a crisis counsellor faster. As a result they are able to respond to 94% of high risk texters in less than 5 minutes (Source)(Source),

AiA Thoughts:

Crisis Textline is a fantastic example of how the collection and analysing of data doesn’t have to be a distraction from your core purpose. Their approach demonstrates how you can successfully use data to enhance the delivery of your service without it being a distraction from what you provide. 

It is also a great demonstration of how a learning culture can be beneficial and that you don’t have to feel pressure to get it right the first time. Their approach to this problem was to take it in stages: they started with a basic list of words to flag their system and then worked to improve it and refine it over time. With any project it can help to start with something simple, see what works, what doesn’t and then use that information to take it from there. A healthy data foundation on which to build can be instrumental in our future successes. 

Predicting demand, MNWD and DataKind

Southern California was facing the worst drought in 500 years and the Moulton Niguel Water District. MNWD worked with DataKind to develop a forecast for the demand for water, which included a range of uncertainty. As a result, MNWD could anticipate peak demand times and plan accordingly so it could reduce the need to import expensive water to supplement supply (Source).

AiA Thoughts:

Again, this is an incredibly inspiring example of data use within the non-profit sector. The accurate forecasting of demand, whether its for water, or any other type of service, can be a struggle for a number of organisations. The desire to have the resources readily available to be directed where it is most needed swiftly is where the priority lies and, at the same time being conscious of stretching that capability as far as it is able to go is vital to the organisation’s longevity. If your organisation also manages a fluctuating need for what it is you provide and yet predicting this need is difficult, then moving your attention to initiating (or refining) demand forecasting may yield similarly promising results. 

Demand forecasting successfully supports non-profits in optimising resource planning. Whether it is designing shift patterns for staff, running workshops etc. it helps ensure that whatever needs to be in operation is there for your beneficiaries when it is needed. 

Predicting high potential donors, Gravyty Technologies

Gravyty uses AI to predict which donors are most likely to make high value donations. This then allows fundraisers to focus their time and resource at those selected individuals, allowing for high degree of personalisation and customisation of outreach activities. Fundraising at Cure Alzheimers Fund increased by 49% after using Gravyty (Source).

AiA Thoughts:

Our final example for this article displays an impressive use of AI in the realm of fundraising. Fundraising is arguably one of the areas which has experienced the most change when it comes to organisations becoming data conscious. This is largely due to the direct and highly visible results it can have on a non-profit’s bottom line. A significant cost base can often be required to generate funds, and often the old adage “we get 80% of our funds from 20% of our effort – we just don’t know which 20%” seems only too often to be an accurate state of affairs. Fundraising can be an excellent place for organisations to look for a data project because:

  • the data itself often already exists within a CRM or accounting software
  • it can have bottom line results, that can inspire further investment into other projects.

We must mention, however, the importance of maintaining balance.  Focusing on fundraising is of obvious importance to non-profits and it can be all too tempting to make this the sole focus of the organisation, at the expense of other areas. Balancing the need to raise funds with looking, as well, at how data can be used to maximise the impact of your organisation ensures a more concrete and long-lasting success moving forward.


Most organisations have a wealth of data across their organisation but often struggle to know where and how to focus their resources into running analytics projects which can deliver the most benefit. From fundraising, to demand modelling, data can be used in all different areas of your organisation to create value.

How can you use data to make your organisation more effective?

Leave a Reply

Your email address will not be published. Required fields are marked *