Want to have a positive social impact as a data scientist?

Part I of a series on social impact for data scientists

As a community, we know we can contribute more than just doctoring startup growth numbers or optimizing ad clicks. But how?

Unfortunately, we don’t actually know yet how to best use data science for making the world a better place. Unlike traditional social impact professionals such as community organizers and fundraisers, we can’t rely on tried career paths and vocational training to help us find our way.

What kind of impact is meaningful to us is a deeply personal question, so there will be no one-fits-all answer. Instead, I want to provide you with some guidance on how to find your own version of meaningful social impact as a data scientist.

To discover what works for us, we need to use what we do best as data scientists: Gather data and learn from it. We can experiment with different approaches to invest our skills for having social impact to the ones that work best for us.

How can you find your personal approach towards positive social impact?

I recommend taking a two-phase path: First, find your general strategic direction by exploring social issues and organizations addressing them. Then, find out the best ways to implement your strategy by experimenting with directions, challenges, and methods and tools.

Before we start on this journey of exploration together, let me briefly introduce Alina. She graduated three years ago with a MSc. in Data Science from the University of Potsdam, Germany. Since then, she has worked as a Junior Data Scientist in a large Berlin-based travel company on pricing strategies. Her work is fun, and she is learning a lot every day. But recently, as she is approaching the magical 30s, she has been wondering how she could experience more meaning in her work. We will follow Alina through this journey.

While the steps of this journey are presented in sequence, you don’t actually have to follow this path in a rigid order. Feel free to skip around if you already have answers to some of these questions.

Just please, don’t start with the tools! This might be tempting because it is what we usually are most comfortable with. But you won’t be effective if you have no clear picture of where you are going.

If you don’t know where to start and want to explore every step, start with the strategic questions. Working your way through this path will be the most helpful to you.

Strategic questions

These questions will help to make sure your experiments are moving into the right overall direction. Before you know where you want to go, optimizing the route to take is not very effective. To answer these strategic questions, you need some time for reflection and research.

1. What social issue do you want to address the most?

E_2018_SDG_Poster_without_UN_emblem_1024px.png

The UN Sustainable Development Goals are a widely used framework for coordinating globally around social issues considered to be most pressing right now.

Which ones are the most relevant to you?

Image copyright: United Nations Department of Public Information, United Nations, S-1018, New York, NY 10017, USA

There are different methods you can use to decide on your most important social issue.

Emotional attraction

Some causes may be closer to your heart than others. Unlike the rational methods, you may not have clear reasons for your choice of cause. But having an emotional connection to a social issue can be a very powerful motivator to sustain your investments.

Alina realizes that she is emotionally affected by the homeless people she meets everyday on her way to work. It saddens her to see them collecting bottles for the deposit, trying selling their newspapers and begging in the metro. So she makes a mental note that this is one of the causes for her to consider.

Random House, 2009

Random House, 2009

Saving as many lives as you can with your available resources

This is a rationality-based method. It focuses on reducing "suffering and death from lack of food, shelter and medical care" as efficiently as possible by investing into impactful and effective aid programs. Proponents of this method argue that especially for donations, this strategy reduces the most human suffering per dollar invested. An extension of that strategy considers reduction of suffering to all living beings, such as animals.

Alina considers this argument. She finds merit in it, but also she realizes that it is difficult for her to feel passionate enough about an issue that she has no direct experience of. And she isn’t convinced she will be able to sustain her engagement in locations so far away that she will never see the effect of her work.

Image: Future of Humanity Institute -  https://www.fhi.ox.ac.uk/press/images/   CC BY-SA 4.0

Image: Future of Humanity Institute - https://www.fhi.ox.ac.uk/press/images/ CC BY-SA 4.0

Reduction of existential risks

Another one of the rationality-based methods is to invest into what is called "Reduction of existential risks". According to its proponents, the most effective impact we can have is by investing our skills towards preventing risks that threaten humanity's long-term survival. Examples of such risks are global warming and nuclear war.

Some of the suggested “existential risks”, such as the threat of hostile AI threatening humanity’s survival seem a bit far-fetched to Alina. However, global warming scares her deeply. So that is another issue that she adds to her list.  

Photo:  Priscilla Du Preez  on Unsplash

Photo: Priscilla Du Preez on Unsplash

Causes that affect you or your local community personally

It can be very powerful to work on issues that you experience yourself. You will both bring a unique perspective as someone affected by the issue. And you may experience a great sense of empowerment from being able to act on something that affects you or your local community rather than just suffering from it.

Alina considers this and she realizes she is pretty privileged and not directly affected by many social issues. On occasion, she has experienced sexism. She considers herself a feminist, so she is adding feminist activism to her list.

Social sector capacity building

Rather than pursuing direct impact, you may want to help the social sector become more impactful overall. Typical examples of such indirect impact are organizations that provide training or technical infrastructure to other social impact organizations. The various “Data for Good” organizations such as “Data Science for Social Good Berlin”, CorrelAid and DataKind fall into this category.

Alina decides that right now, she prefers to engage directly with social issues rather than capacity building.

Out of these methods, decide which one you prefer. And then pick one or two issues that you want to explore further.

For Alina, those are homelessness and climate change. While feminist activism is something she considers meaningful as well, she currently finds that issue less pressing than the other two.

2. Which organizations do you want to invest your efforts into?

Now that you found your most meaningful social issues, it is time for some research! You want to compile a list of organizations that are working on these issues, ideally that are based geographically close to you.

For the organizations on your list, you want to find out whether they are achieving what you want to invest your efforts into. Find and review their annual reports, search whether they outline how they expect their programs to have an impact and whether they publish evidence for the impact of their work. For larger organizations, search for any news mentions as well.

Out of your research, select a few organizations (I would recommend less than 5) you are interested in investing your efforts into. If you can't find any organizations that suit you, it might also be worth investigating whether you want to start your own organization.

Photo: Matt Collamer on Unsplash

Photo: Matt Collamer on Unsplash

After some research, Alina found ten interesting Berlin-based organizations, five working against global warming and five working with homeless people. Out of these ten, she could find some evidence for positive impact for seven of them. From those, she decided that these four are the most motivating organizations for her:

  • A small nonprofit providing on-the-street services such as food, clothing and sleeping bags to homeless people

  • An advocacy organization trying to encourage IT companies to switch their servers to 100% green energy

  • A political party pursuing regulatory policies she considers to be more effective in both of her chosen issues

  • An activist organization helping to organize the local “Fridays for Future” climate protest marches

3. Which impact do you want to achieve with your investments?

Since you know what issues you want to work on, you can conduct some more research to understand the current state of the issue: How severe is it? Who is most affected? What is the scope of it? Where is it most prevalent?

Based on these insights, you can give your impact ambitions a clearer direction. When would you consider your investments to yield an impact that feels meaningful? Again, this is a very personal question.

Try to define your ambitions as precisely as possible. Who are the people whose situation should have improved? By what amount? How many people do you want to impact? To find out later whether you are achieving the impact you want, create some concrete goals now.

This has been a tough question for Alina. Her chosen causes and four chosen organizations are very different, so it takes her some thinking to figure out her impact goals. One thing she realizes is that she experiences immediately tangible impact as more meaningful than intangible impact such as advocacy.

She sets her primary goal around helping homeless people in Berlin. Googling statistics tells her that there are estimated to be between 6000 and 10000 homeless people in Berlin. As her own ambition, she decides that helping around 300 of those homeless people within the next year would feel like a meaningful impact.

Implementation questions

Now that you know which social issues you want to address, which organizations you want to invest in and which impact you hope to achieve, you are ready to conduct some experiments! You can experiment to find which path is the best to implement your strategy.

Your experiments help you to address the following implementation questions

1. Which direction do you want to take in your work with your chosen organizations?

2. Which challenges do you want to address that your organization faces?

3.    Which data science methods and tools do you want to use?

overall_method.png

In a future blog post, I will suggest some more detailed ideas how you can conduct your experiments. For now, let’s keep going with Alina’s story.

1. Experiment with the direction you want to take

There are many ways to work with organizations to have a positive impact. Each of these directions works for different resource constraints of your organization. You can make a data science contribution in any of these directions. Some of the options you can experiment with are

Photo:   William Iven     at  Unsplash

Photo: William Iven at Unsplash

  • Volunteering your skills: Works well for nonprofit organizations that are small or implementing small projects in large nonprofit organizations. As a data scientist, you can contribute a unique skillset to your chosen organizations.

  • Improving ethical standards in your industry: Works well when you work in industry, have some informal leadership experience and your organization lacks expertise on implementing existing ethical best practices. There are plenty of opportunities to improve ethical standards for data science work in any industry.

  • Academic research: Works well where no current best practices exist yet to address the social issue. Academic research often requires conducting and analyzing empirical studies, data scientists can help with that.  

  • Donating part of your income: Works well if your selected organization is more constrained by funding than staff or knowledge limits. Data science as a high-earning career is a great base for donating a part of your income.

  • Advocating for regulatory policy changes: Works well when best practices exist that need to be implemented on a large scale across many organizations.  Certain types of advocacy work can benefit from data science methods such as target group segmentation, A/B testing and impact measurement.

  • Professional non-profit work: Works well if best practices exist that need high effort and long-term implementation projects. Also in this situation, as a data scientist you can add a special skill set to your organization of choice.

There are probably many more that I am forgetting. Feel free to reach out to me and suggest more!

A rule-of-thumb for selecting the directions you want to experiment with is to learn where your chosen organization faces the most resource constraints.

Alina was speaking with the staff of the small nonprofit providing on-the-street services such as food, clothing and sleeping bags to homeless people. She has learned that they receive sufficient funding for their work, but are limited in their impact by lack of staff, especially volunteer staff. They also mentioned that their work would be much easier if some local laws were changed into being more favorable to their efforts.

Therefore, Alina decides to experiment with volunteering for this organization and also joining the political party she evaluated earlier in order to advocate for more helpful local laws. Joining that party could be considered both a donation in the shape of membership fees and an effort in advocating.

2. Experiment with challenges to address

Organizations face different challenges when working to address social impact. A good understanding of your organization's work and current situation will help you to pick a few to experiment with. Work with your organization's staff to figure out where the current bottlenecks are!


Some of the options for challenges that your organization might face:

  • Implementing their programs

  • Fundraising

  • Advocacy

  • Organizational management

  • Staffing

  • Monitoring impact

There are definitely more options here that depend on the specific organization. Your organization's staff will know what those are.

Alina is pretty excited to get involved with her new organization and party. She is wondering what it will take specifically for he to achieve her goal of helping 300 homeless people this year.

So she asks that question to Robert, the execute director of the small nonprofit. “I don’t have an answer to that question right now”, Robert told her. “We don’t track our impact to that level of detail.” “Well, I can help you find that out!”, replies Alina.

Alina now knows that monitoring impact is one of the bottlenecks of her organization. How can they maintain and improve their level of service if they don’t know what they are achieving right now? Maybe having clear evidence of their impact could also be helpful in advocating for improved laws? She decides that this is an experiment she wants to try out.

3. Experiment with data science methods and tools

This is the stuff you are probably already familiar with, so I won't go into too much detail.

One important thing to figure out before you experiment is to find out which tools and methods fit within the context of the organizations you want to work with.

You may have to work with less effective tools depending on what the organization can offer. For example, you may have to rely on whatever licenses big tech companies donate through programs such as TechSoup rather than being able to choose your own tools.

Photo:  Roven Images  on Unsplash

Photo: Roven Images on Unsplash

Alina is happy now. She now knows what she wants to experiment with: She will volunteer for her chosen nonprofit for the next year and help them measure how many homeless people they are impacting and how. She is hoping that within this year, she will allow them to see some significant improvements in those numbers and also present their impact to local lawmakers to advocate for more favorable laws. Have fun, Alina, we will check in with you again after one year!

Most importantly: Have fun!

Learning about yourself and making a difference should feel fun, inspiring and meaningful.

So, don't worry if these are a lot of steps. Don't worry if it feels overwhelming right now. Take your time! I have been experimenting with my approach to social impact for over 15 years, and I am still learning more every day. I hope you will, too!

Special thanks to: Lisa, Andrew and Kat for their reviews and feedback on this post!