Data Science for Social Good: 2024 Trends and Best Practices

Data science offers powerful tools to understand and address some of the world's most pressing problems. When people use these tools to make a positive difference in society, it is called Data Science for Social Good (DSSG). This field has grown significantly, with new methods and applications emerging constantly. This report explores the current state of DSSG, highlighting recent advancements, key technological trends, important challenges, and guidelines for responsible practice.

The idea of using data to benefit humanity is not new, but the tools and approaches have become much more powerful. Today, Data Science for Social Good (DSSG) is an active and expanding field.

In 2023 and 2024, DSSG means using data analysis skills and computational methods to help solve important societal problems. This work covers a wide range of issues, including health, education, human rights, economic opportunity, and environmental protection. The core idea is to apply the power of data to achieve positive social impact.

Many universities and organizations now train people specifically for this work. For example, Stanford University's Data Science for Social Good (DSSG) summer program brings together aspiring researchers to work on projects with governments and non-profit organizations. These projects tackle real-world problems in areas like education, health, public safety, and international development. Similarly, the University of Chicago offers a Data Science for Social Impact (DSSI) Summer Experience. This 8-week paid research program for undergraduate students focuses on research and data science to make a positive social impact, covering topics such as climate, health, policy, and human rights. JPMorgan Chase also supports students in this area through hackathons where they use tech skills to solve real-world problems for social good organizations.

The ultimate goal of these efforts is to make a real, measurable difference in people's lives and communities. This focus on tangible outcomes is a defining feature of modern DSSG. The increasing number and variety of these structured programs, from summer schools to strategic government initiatives like those outlined by the National Institutes of Health (NIH) for training, show that DSSG is moving from a niche interest to a recognized field. This formalization creates clearer pathways for individuals to get involved and helps build a skilled workforce dedicated to social impact, suggesting a growing demand from various organizations for these specialized skills.

Why using data thoughtfully helps communities

Thoughtful use of data provides many benefits to communities. Data can clearly show where problems are most severe and which groups of people need the most help. This allows organizations, which often have limited money and staff, to use their resources more effectively. For instance, data analysis can help target poverty reduction programs to the specific areas where they will achieve the greatest impact, ensuring that aid reaches those who need it most.

Data also helps us understand complex issues more deeply. A 2024 Stanford DSSG project, for example, studied food reactivity in children as reported by parents. The project aimed to identify patterns, understand the severity of reactions, and assess the social implications, providing valuable insights for parents and healthcare providers. Another example comes from TREND Community, which uses artificial intelligence (AI) to analyze conversations on social media. This helps them understand the challenges faced by people living with rare and chronic diseases, such as identifying unmet needs or patterns in how diseases present themselves.

Crucially, when communities themselves participate in how data is collected and used, the solutions developed are more likely to be effective and fair. TREND Community actively listens to community conversations to guide its work. The University of Chicago's Data Science Institute also emphasizes community-centered data science in its projects. The Community Planning + Visualization Lab at Rowan University provides another example, using community-engaged research methods, citizen science initiatives, and community advisory boards in its projects. This participatory approach ensures that projects align with the actual needs and contexts of the people they intend to serve, fostering trust and a sense of ownership within the community. This shift towards community involvement reflects a learning process within the DSSG field: outcomes are better when the people impacted are part of the solution-finding process.

The range of issues DSSG now addresses is also expanding. Early projects might have focused on specific local social services. However, recent examples show applications tackling global issues like climate change, human trafficking, and complex health concerns. The inclusion of environmental justice and difficult international problems demonstrates that DSSG is becoming a versatile tool applicable to almost any area where data can offer insights for improvement. This reflects a more complete understanding of what contributes to societal well-being.

Guideline: Writing clearly about complex topics

To explain complex ideas effectively, use simple language. Short sentences are often easier to understand than long ones. Try to avoid technical words, or jargon. If a technical term is necessary, explain its meaning immediately.

For example, instead of writing: "DSSG leverages advanced computational methodologies to derive actionable insights from heterogeneous datasets," one could say: "Data science for social good uses powerful computer methods to study different kinds of information. This helps us understand problems and find solutions." Following this guideline is very important for making information accessible, especially when aiming for an 8th-grade reading level.

New Ways Data Science is Making a Difference (2023-2024 Highlights)

Recent years have seen exciting new applications of data science that directly benefit people, societies, and the environment. These projects often combine different data science techniques to achieve more powerful results and are increasingly used to hold institutions accountable and to anticipate future needs.

Helping People and Society

Improving health access and understanding

Data science plays a crucial role in addressing public health concerns. For instance, a 2024 Stanford Data Science for Social Good (DSSG) project focused on understanding parent-reported food reactivity in children. The study aimed to identify patterns, determine the severity of reactions, and assess the impact of dietary changes. Such research can lead to better support systems for affected children and families.

The National Institutes of Health (NIH) also recognizes the power of data in health. Their Strategic Plan for Data Science for 2023-2028 aims to use data to better understand health across all populations and to encourage the ethical use of artificial intelligence (AI) in healthcare. A key part of this plan involves improving how human-derived data, including clinical records and real-world information, is used for research. The NIH also seeks to better understand how social and environmental factors contribute to health equity.

Organizations like TREND Community use AI to analyze public conversations on social media about rare and chronic diseases. In 2023, their work on Sjögren's disease helped to confirm that "flares" are a significant part of the patient experience and identified common co-occurring symptoms. These findings can improve patient care and guide further research.

Geospatial analytics, which involves using location-based data, is another powerful tool. It helps in coordinating the delivery of humanitarian aid and in understanding the social factors that affect community health. News stories from Esri's GIS for Good initiative between 2023 and 2025 highlight practical applications. For example, Public Health Scotland uses location data to improve healthcare services, and cities use mapping to identify communities most vulnerable to extreme heat. These examples demonstrate data science moving beyond broad statistics to address specific health challenges, ultimately improving well-being through better understanding and resource allocation.

Protecting human rights and promoting justice

Data science offers powerful tools to uncover injustices and protect vulnerable people. A 2024 Stanford DSSG project provides a compelling example. Working with the Stanford Human Trafficking Data Lab and the Brazil Federal Labor Prosecutor Office, student fellows developed a machine learning solution to detect illegal charcoal production sites in Brazil from satellite imagery. These sites are often linked to human trafficking, forced labor, and illegal deforestation in the Amazon rainforest. The new model significantly reduces the high number of false alarms from a previous system, allowing human investigators to focus their efforts more effectively and improving the chances of successful interventions against labor exploitation.

Investigative journalism also benefits from data science. The "Missing in Chicago" series by the Invisible Institute, which received a 2024 Pulitzer Prize, used data-based inquiry to examine how the Chicago Police Department handles cases of missing persons, particularly Black girls and women. Trina Reynolds-Tyler, a data director at the Invisible Institute, mentioned using a classifier algorithm in a related project called "Beneath the Surface." This algorithm helped identify instances of police misconduct related to gender-based violence and missing persons cases by analyzing police complaint records. These projects show data science becoming a tool for civic action and advocacy, enabling communities and journalists to demand transparency and reform from powerful institutions. This can lead to systemic change rather than just isolated interventions.

Organizations like LIRNEasia are also working to promote AI for social good. In 2023 and 2024, they held workshops to raise awareness about AI's potential to benefit society, including in areas related to justice and human rights.

Making education more fair

Education is fundamental to opportunity, and data science can help make educational systems more equitable. Predictive analytics can analyze student data—such as attendance, performance, and socioeconomic background—to identify students at risk of falling behind academically. This early identification allows educators to offer personalized support and interventions, which can reduce dropout rates and improve overall academic success.

The University of Chicago's Data Science for Social Impact (DSSI) Summer Experience includes projects that touch on policy and human rights, areas that often intersect with education equity. Furthermore, presentations on platforms like SlideShare discuss how DSSG training programs affect student experiences and meet workforce demands. They also show how universities use data to address issues like student attrition and to create more personalized learning environments.

Mapping technologies also play a role. They can guide the development of culturally responsive STEM (Science, Technology, Engineering, and Math) education, ensuring that teaching approaches are sensitive and relevant to diverse cultural contexts. An Esri GIS for Good story from February 2024 highlighted this application.

Protecting Our Planet

Using maps and location data (Geospatial Analytics) for environmental justice and conservation

Geospatial analytics, the analysis of data linked to specific locations, is a powerful tool for understanding and addressing environmental issues. As the global urban population continues to grow, this type of analysis is critical for monitoring urban expansion, optimizing land use, and improving the quality of life in cities. The market for these tools is expanding, with strong demand from sectors like transportation, government, and healthcare. A notable development in October 2023 was the launch of SoilMate.ai, a web-based service focused on remote earth analytics.

Esri's GIS for Good initiative regularly features stories from 2023-2025 showcasing how these tools make a difference:

In September 2024, a new mapping tool was highlighted that helps guide New York City's efforts toward environmental justice.
An August 2023 story detailed how Los Angeles County uses mapping to ensure equitable access to natural spaces and to guide the remediation of degraded lands.
Also in August 2023, the U.S. Centers for Disease Control and Prevention (CDC) released a new map that allows people to see levels of environmental injustice in their communities.
In May 2023, the Environmental Protection Agency (EPA) was featured for using maps of Superfund (heavily polluted) sites to inspire and guide community revitalization efforts.

The Stanford DSSG project mentioned earlier, focused on detecting illegal charcoal production in Brazil, also directly addresses environmental concerns by helping to combat illegal deforestation in the Amazon. Location data provides a powerful visual and analytical means to understand environmental problems, identify disparities, and plan effective solutions for a healthier planet and fairer communities.

Addressing climate change impacts

Climate change is a global crisis, and data science is essential for understanding its complex dynamics and developing strategies to mitigate its effects. Data scientists develop sophisticated models that analyze historical climate data, satellite imagery, and atmospheric measurements to forecast long-term climate trends. These forecasts help scientists and policymakers anticipate future changes and plan accordingly.

Satellite imagery and remote sensing technologies are also used to monitor deforestation and the destruction of natural habitats in real-time. Data scientists process this information to create detailed maps showing changes in forest cover, which helps conservationists and governments target their protection efforts more effectively.

The Future Today Institute's 2024 report observed that as the effects of climate change become more severe, governments in 2023 gave serious consideration to more unconventional technologies like solar geoengineering and ocean chemistry manipulation to combat the situation. The report also noted significant progress in enabling infrastructure for renewable energy, with a focus on smart grids, better energy storage solutions, and improved carbon tracking methods.

Academic initiatives also contribute. The University of Chicago's AICE (AI for Climate) initiative works to integrate AI with fundamental domain knowledge. The goal is to transform climate research, focusing on both scientific advancements and their societal impacts. These efforts show a trend towards more proactive and predictive applications of DSSG. Rather than only analyzing past events, data science is increasingly used to forecast future scenarios, such as identifying at-risk students before they struggle or predicting climate trends to enable better preparedness. This shift allows for preventative measures, which can reduce harm and save resources compared to reacting after a problem has fully emerged.

Recent DSSG Projects Overview

Project/Initiative Name	Area of Impact	Key Data Science Method/Technology Used	Brief Description of Outcome/Goal
Stanford DSSG: Detecting Illegal Charcoal Sites	Human Rights, Environment	Machine Learning, Satellite Imagery Analysis	Reduced false positives from CNN model, improved efficiency of field inspections against labor exploitation and deforestation
Stanford DSSG: Food Reactivity in Children	Public Health	Data Analysis, Dietary Modification Study	Assessed prevalence/severity of food reactivity, explored diet-symptom relationship
TREND Community: Sjögren's Disease Flares	Healthcare (Rare Diseases)	AI, Social Media Listening, NLP	Confirmed flares as salient, identified co-occurring symptoms (pain, dryness, fatigue), findings presented at conference
Esri/NYC: Environmental Justice Mapping	Environmental Justice	Geospatial Analytics, Mapping Tools	Guiding New York City towards environmental justice
Invisible Institute: Missing in Chicago	Justice, Human Rights	Data-based inquiry, Classifier Algorithm	Revealed issues in police handling of missing Black girls/women cases, won Pulitzer Prize
NIH Strategic Plan: Predictive Health Models	Public Health, Health Equity	AI/ML, Predictive Analytics, Data Harmonization	Develop predictive models for health outcomes, enhance understanding of social determinants of health
Microsoft Aurora AI: Global Weather & Pollution	Climate, Environment	AI Foundation Model, Predictive Modeling	Predicts global weather patterns and air pollution with high accuracy, operates much faster than traditional systems

Guideline: Using active voice to describe actions and impacts

When writing, focus on who or what is performing the action. For example, instead of saying, "Resources were optimized by data science," say, "Data science optimized resources." This makes sentences clearer and more direct, which helps in maintaining an accessible reading level.

Technological advancements continually provide new tools and opportunities for Data Science for Social Good. Artificial intelligence, geospatial analytics, and the contributions of citizen data scientists are particularly noteworthy trends.

Artificial Intelligence (AI) for Good

Artificial intelligence is rapidly evolving, offering more sophisticated ways to address complex social issues. The Future Today Institute's 2024 report identifies AI as a transformative technology. Key advancements include multi-modal AI, which can process and understand different types of data simultaneously (like text, images, and sound), and self-improving AI agents that can learn and adapt over time. These capabilities can significantly change how we approach social challenges.

Generative AI, a type of AI that can create new content such as text, images, or even music, shows great promise for social good. It has potential applications in areas like personalized healthcare, creating tailored educational materials, assisting in disaster response efforts, and supporting humanitarian aid. For example, generative AI could help develop new medical imaging techniques or create communication aids for diverse populations.

As AI systems make more decisions that affect people's lives, the need for Explainable AI (XAI) becomes increasingly important. XAI aims to make the decision-making processes of AI models transparent and understandable, even to non-experts. This is especially critical in sensitive areas such as healthcare, where AI might assist in diagnoses, or in the justice system, where AI could be used to assess risk. Research in 2023 and 2024 has focused on developing methods to interpret AI models, for instance, in applications like managing water quality, to ensure that the decisions made by AI are clear and justifiable.

AI also plays a role in combating misinformation. Researchers are developing AI systems to detect and flag fake news and disinformation, which is vital for maintaining healthy and informed societies. The increasing accessibility of AI tools, while empowering, also brings risks. For example, the wide availability of generative AI like ChatGPT means these tools could be misused, such as to spread false information, if not paired with strong ethical training and oversight. This makes the development of XAI and adherence to ethical AI principles even more critical, especially as individuals who are not AI experts begin to use these powerful tools.

AI in healthcare and public health advancements

AI is making significant contributions to healthcare and public health. The NIH Strategic Plan for Data Science (2023-2028) places a strong emphasis on developing "trustable AI." This means creating AI systems that are fair, reduce bias, minimize risks, and are explainable, transparent, and adhere to FAIR data principles (Findable, Accessible, Interoperable, and Reusable).

In practical terms, AI helps optimize the allocation of healthcare resources, such as staff and equipment. It can identify communities most in need of health services and ultimately improve patient outcomes. TREND Community provides a specific example of AI in action. They use a proprietary AI analytics engine called Krystie™, which employs machine learning and natural language processing, to analyze conversations among patients on social media. This allows them to gain deep insights into various diseases, such as understanding the patient experience with Sjögren's syndrome.

Academic institutions are also driving innovation. In 2024, Brown University's Data Science Institute awarded seed grants for projects exploring AI's role in public health, including one titled "Leveraging Artificial Intelligence and Data Science to Improve Community-Based and Public Health Initiatives".

A promising AI technique called federated learning is also gaining traction in healthcare. Federated learning allows multiple institutions to collaboratively train AI models on their combined health data without actually sharing the raw patient data. This method enhances patient privacy while still allowing for the development of robust AI models. Studies have shown that federated learning has improved results in areas like breast density classification, predicting outcomes for COVID-19 patients, and accelerating drug discovery processes.

Using Location Data (Geospatial Analytics): New tools and their impact

Geospatial analytics, which focuses on data tied to specific geographic locations, is becoming increasingly vital for social good. It is critical for effective urban planning, helping cities manage growth, allocate resources efficiently, and improve the overall quality of life for residents. The demand for geospatial solutions is rising across various sectors, including transportation, defense, government, and healthcare.

A key trend is the integration of AI and machine learning (ML) with geospatial data. This combination significantly enhances the ability to extract actionable insights from location-based information. An example of this is SoilMate.ai, a web service launched in October 2023 that provides remote earth analytics, likely using AI to interpret satellite or other geographic data.

Esri's "GIS for Good" initiative consistently highlights innovative applications of geospatial technology in 2023-2025:

Demining with Drones (October 2024): Using drones equipped with sensors and analytical software to locate and map landmines, making dangerous areas safer.
Environmental Justice in NYC (September 2024): A new mapping tool helps guide New York City's efforts to achieve environmental justice by identifying areas disproportionately affected by pollution or lacking green spaces.
Detailed Flood Risk Mapping in LA (August 2023): Advanced mapping techniques reveal hidden flood risks in Los Angeles with unprecedented detail, improving preparedness.
Reconnecting Isolated Communities (April 2023): Maps help communities previously cut off by highways to plan new futures, addressing historical inequities.

The Stanford DSSG project on illegal charcoal production also effectively uses geospatial data. The project incorporates geospatial covariates (location-based variables) along with census and survey data to train and improve its machine learning model for detecting illegal sites. Location is a fundamental aspect of many social and environmental problems. Geospatial tools, especially when combined with AI, offer powerful ways to visualize, analyze, and ultimately address these complex issues.

The Role of Citizen Data Scientists: How everyday people contribute

The rise of citizen data science is a significant trend making data analysis more accessible. Citizen data scientists are often individuals who are not formally trained as data scientists but use user-friendly data tools and "augmented analytics" (AI-assisted data analysis) to gain insights and contribute to projects.

Government platforms like CitizenScience.gov in the U.S. provide listings of projects where volunteers can participate in scientific research. For instance, the Community Collaborative Rain, Hail & Snow (CoCoRaHS) network has a Data Explorer tool, launched in April 2024, that empowers citizen scientists to contribute and access weather data.

The Civic Tech Field Guide lists numerous organizations that facilitate citizen involvement in technology for social good. Examples include "Coders Beyond Borders," which empowers newcomers by teaching them tech skills, and "Data For Crisis," which uses data visualization to support disaster response efforts, often relying on contributions from a broad community.

Educational programs also foster early-career citizen data science. The University of Chicago's DSSI Summer Experience, for example, involves undergraduate students working directly on social impact projects, applying data science skills to real-world problems. Citizen data science democratizes the process of data analysis. It allows more people and communities to participate in understanding and solving problems that directly affect them. This not only expands the capacity for social good work but also brings diverse perspectives to the problem-solving process.

Guideline: Explaining technical trends in simple terms

When discussing complex technologies, break them down into their basic functions and benefits. Use analogies if they can help clarify the concept. Focus on what the technology does for social good, rather than just providing a technical definition of what it is.

For example: "Multi-modal AI means AI that can understand different types of information at once, like words, pictures, and sounds. This is similar to how a person can read a story, look at the illustrations, and listen to a narrator all at the same time to understand it better. For social good, this could help doctors by enabling an AI to analyze medical images while also reading patient notes to help make a more accurate diagnosis."

Important Challenges We Must Address

While data science offers tremendous potential for social good, several significant challenges must be addressed to ensure its responsible and effective use. These challenges range from ethical considerations in data handling to practical issues of data quality and resource availability.

Making Sure Data is Used Fairly and Ethically

Avoiding bias in data and algorithms

One of the most critical challenges is ensuring that data and the algorithms trained on them are free from unfair bias. Machine learning algorithms learn from the data they are given. If this data reflects existing societal biases (for example, historical discrimination in loan applications or hiring), the algorithm can inadvertently learn and even amplify these biases. This is a major concern highlighted in the NIH Strategic Plan for Data Science, which explicitly aims to develop programs to reduce unintended biases in AI and ensure that AI is FAIR (Findable, Accessible, Interoperable, and Reusable), validated, and explainable.

It is essential to design and implement AI systems that do not discriminate against certain groups of people. Addressing this requires several actions: implementing technical fairness measures during algorithm development, regularly auditing algorithms for biased outcomes, and providing comprehensive fairness-awareness training for data scientists and developers. UNESCO's Recommendation on the Ethics of AI also strongly emphasizes the principles of fairness and non-discrimination, urging AI actors to promote social justice and ensure that AI's benefits are accessible to all. If AI systems perpetuate bias, they can lead to unfair and harmful outcomes in critical areas like loan approvals, job opportunities, healthcare access, or even criminal justice. This would undermine the very purpose of "social good" and could disproportionately harm already marginalized communities.

Protecting people's privacy and data

The increasing collection and use of personal data in DSSG projects make privacy and data security paramount. The NIH strategic plan, for instance, prioritizes the protection of research participant privacy in all its data science initiatives.

Obtaining truly informed consent from individuals whose data is used can be complex, particularly when dealing with vulnerable populations or when data is combined from multiple sources. It is crucial that people understand how their data will be used, what the potential risks are, and that they have clear options to opt out if they choose. UNESCO's ethical principles also highlight the fundamental right to privacy and the need for adequate data protection frameworks throughout the entire lifecycle of an AI system. Breaches of privacy or misuse of personal data can lead to serious consequences, including identity theft, discrimination, financial loss, or emotional distress. Maintaining strong data protection practices is essential for building and maintaining public trust, which is a cornerstone of successful and ethical DSSG projects.

Being open about how data is used (transparency and accountability)

Transparency in data science means that people should be able to understand, at an appropriate level, how data science methodologies work and how findings are derived, even if they are not technical experts. This involves providing clear explanations of the methods used, the assumptions made, and how AI models arrive at their outcomes.

Accountability means that data scientists and the organizations they work for must take responsibility for the consequences of their work. This includes any unintended negative impacts that may arise from the use of data science or AI systems. UNESCO's principles call for AI systems to be auditable and traceable, with mechanisms for oversight and impact assessment. The NIH plan also emphasizes the need for transparency in AI. Transparency builds trust and allows for public scrutiny and independent review of DSSG projects. Accountability ensures that if errors occur or harm is caused, there are established processes for addressing these issues, learning from mistakes, and making amends where necessary. Without clear lines of responsibility, particularly when complex AI systems are involved, it can be difficult to determine why an error occurred or who is at fault. This "accountability gap" can undermine trust and the ethical foundation of DSSG.

Data Quality and Access

The quality and accessibility of data are fundamental to the success of any DSSG project. Poor-quality data—data that is inaccurate, incomplete, or outdated—can lead to incorrect conclusions and, consequently, flawed decisions or interventions. Ensuring the accuracy and reliability of data is an ongoing challenge that requires careful data collection, cleaning, and validation processes.

The NIH Strategic Plan aims to improve capabilities for data management and sharing across the biomedical research community. A key part of this is promoting FAIR data principles, which advocate for making data Findable, Accessible, Interoperable, and Reusable. The plan also emphasizes the need for greater data harmonization, which means ensuring that data from different sources can be combined and compared effectively.

Another significant challenge is the "digital divide." This term refers to the unequal access to digital technologies, including the internet and data-driven solutions. If not addressed, the benefits of DSSG may not reach everyone equally and could even exacerbate existing inequalities. For example, if a health intervention relies on smartphone data, it might miss populations without access to such technology. Therefore, efforts must be made to ensure that data science applications are designed to be inclusive and beneficial to all segments of society. Without good quality, representative data, DSSG projects can fail to achieve their goals or, worse, cause unintended harm. Equitable access to data and the benefits derived from it is crucial for truly achieving social good.

Having Enough Resources: People, money, and tools

Many social impact projects, especially those run by non-profits or community organizations, operate with limited resources. This includes constraints on funding, access to skilled personnel, and the necessary technological tools and infrastructure. These limitations can significantly affect the depth, scope, and scale of the data analysis that can be performed, potentially hindering the impact of DSSG initiatives.

There is a well-recognized need for more trained data scientists, particularly those from diverse backgrounds who can bring varied perspectives to social problems. The NIH strategic plan, for example, includes objectives to increase training opportunities in data science and to expand and diversify the data science workforce. Investing in human capital—by training more data scientists and upskilling existing professionals in social sector organizations—as well as providing access to appropriate tools and funding, is essential for the continued growth and effectiveness of the DSSG field.

Environmental Impact of Data Science

An often-overlooked challenge is the environmental footprint of data science itself. Large-scale data analysis, and particularly the training and deployment of complex AI models like deep learning networks, can require substantial computational power. This, in turn, consumes a significant amount of energy, which can contribute to carbon emissions and other environmental impacts, especially if the energy sources are not renewable.

It is important for the DSSG community to be mindful of this environmental cost. This includes considering the energy efficiency of algorithms and computing infrastructure, exploring options for using renewable energy sources to power data centers, and generally promoting sustainable computing practices. The goal of DSSG is to solve societal problems, not to create new environmental ones. Therefore, managing the environmental impact of data science operations is crucial to ensure that the field is truly contributing to overall global well-being.

Key Ethical Challenges Summary

Challenge	Brief Description	Potential Negative Impact if Unaddressed	Recommended Approach/Guideline
Algorithmic Bias	AI models learn and may amplify biases present in data, leading to unfair decisions	Discrimination, perpetuation of societal inequalities, harm to vulnerable groups	Implement fairness measures, audit algorithms, use diverse datasets, provide fairness-awareness training, utilize XAI techniques
Data Privacy	Collection and use of sensitive personal information risks exposure or misuse	Identity theft, discrimination, loss of trust, misuse of data	Employ robust data security, encryption, informed consent, privacy-enhancing technologies, clear data usage policies
Lack of Transparency	"Black box" AI models where decision-making processes are unclear to users/stakeholders	Difficulty in detecting errors/bias, lack of trust, inability to hold systems accountable	Utilize XAI techniques, provide clear documentation of methods and assumptions, use tools like model cards
Data Quality & Representativeness	Use of inaccurate, incomplete, or non-diverse data for training models	Incorrect conclusions, biased outcomes, solutions that do not work for everyone	Implement data validation, Quality Assurance/Quality Control processes, make efforts to collect diverse data, apply FAIR data principles
Accountability Gap	Difficulty in assigning responsibility when AI systems cause harm or error	Lack of redress for victims, erosion of trust, perpetuation of harmful system	Establish clear accountability frameworks, ensure human oversight, make systems auditable and traceable

Guideline: Clearly stating problems and why they matter

When discussing a challenge, explain what the challenge is in simple terms. Then, briefly explain the negative consequences that could occur if the challenge is not addressed. This helps readers understand the urgency and importance of finding solutions.

To ensure that data science genuinely contributes to social good, practitioners and organizations must adopt responsible practices. This involves putting people first, adhering to ethical principles, sharing data safely, and fostering an organizational culture that supports these efforts.

Putting People First: Focusing on community needs and involvement

The most effective Data Science for Social Good projects are those that are deeply connected to the communities they aim to serve. It is vital to truly understand the needs, perspectives, and existing strengths of a community before designing any data-driven solution. TREND Community, an organization that uses AI to understand rare and chronic diseases, emphasizes that "too many voices are left out of the conversation." Their approach involves analyzing community conversations on social media to gain a deeper understanding of patient experiences, disease presentation, and unmet needs. They then empower these communities to advocate for themselves using the evidence-based data and insights generated.

Involving community members directly in the design, implementation, and evaluation of projects is crucial. The Community Planning + Visualization Lab at Rowan University provides a strong example of this. They utilize community-engaged research methods, citizen science initiatives where community members participate in data collection or analysis, and community advisory boards to guide their work. This collaborative approach helps ensure that the solutions developed are relevant to the community's actual needs, are culturally appropriate, and are more likely to be adopted and sustained. Similarly, the University of Chicago's Data Science Institute highlights the importance of community-centered data science through initiatives like its PalmWatch project (a tool to monitor palm oil production) and its Community Data Fellows program, which pairs students with local organizations.

Guideline: Start by listening

Before designing any data solution or intervention, engage deeply with the community. Understand their perspectives, priorities, existing resources, and concerns. Ensure that their voices and insights guide every stage of the project, from initial conception to final deployment and evaluation.

This people-first approach is essential because solutions designed without community input (top-down solutions) often fail to address the real issues or may even have unintended negative consequences. Genuine community involvement leads to more effective, sustainable, and equitable outcomes.

Key Principles for Ethical AI: Learning from global standards

Several international organizations have developed principles and frameworks to guide the ethical development and deployment of Artificial Intelligence. These provide a valuable foundation for DSSG projects.

UNESCO's Recommendation on the Ethics of AI

UNESCO's framework is centered on four core values:

Respect for human rights and human dignity
Environment and ecosystem flourishing
Ensuring diversity and inclusiveness
Living in peaceful, just, and interconnected societies

From these values, UNESCO outlines ten key operational principles for ethical AI:

Proportionality and Do No Harm: AI use must be necessary for a legitimate aim, and risks assessed to prevent harm.
Safety and Security: AI systems should be safe from unintended harm and secure from attacks.
Fairness and Non-Discrimination: AI actors must promote social justice and ensure AI benefits are accessible to all.
Sustainability: AI's impact on sustainability goals (like the UN SDGs) must be assessed.
Right to Privacy and Data Protection: Privacy must be protected throughout the AI lifecycle.
Human Oversight and Determination: Humans must retain ultimate responsibility and accountability.
Transparency and Explainability: AI systems should be transparent and explainable appropriate to the context.
Responsibility and Accountability: AI systems should be auditable and traceable, with oversight mechanisms.
Awareness and Literacy: Public understanding of AI and data should be promoted.
Multi-stakeholder and Adaptive Governance & Collaboration: Inclusive participation is needed for AI governance.

OECD AI Principles (updated 2024)

The Organisation for Economic Co-operation and Development (OECD) also provides influential AI principles. These principles, first adopted in 2019 and updated in 2024, aim to promote AI that is innovative and trustworthy, and that respects human rights and democratic values. The five values-based principles are:

Inclusive growth, sustainable development and well-being
Human-centred values and fairness
Transparency and explainability
Robustness, security and safety
Accountability

The OECD.AI Policy Observatory supports these principles by providing tools, data, and analysis to policymakers and AI actors, helping them to address AI risks such as bias, privacy infringements, and security threats in various contexts, including social good initiatives.

Guideline: Familiarize your team

Familiarize your team with these international ethical frameworks from UNESCO and the OECD. Use their principles as a checklist and a practical guide when planning, developing, and deploying any DSSG project.

The National Institutes of Health (NIH) Strategic Plan for Data Science (2023-2028) strongly advocates for data sharing to maximize the value of research investments, but it equally emphasizes the need for crucial safeguards to protect participants and ensure ethical conduct.

Key elements for safe and effective data sharing, based on the NIH approach, include:

Promote FAIR Data: Data should be managed and shared in a way that makes it Findable, Accessible, Interoperable, and Reusable. This involves creating good data management plans, using clear and comprehensive metadata (data about the data), and adopting common data standards and formats where possible.
Protect Participants: The rights and welfare of individuals who provide data must always be the top priority. This means obtaining truly informed consent, especially when dealing with sensitive information, linking datasets from different sources, or working with vulnerable populations. Ethical frameworks must guide data linkage activities, and participants should understand how their data will be used and have control over their participation.
Implement Strong Data Governance: Data repositories and sharing platforms must have trustworthy governance structures. This includes clear policies for data access, use, and security. For specific communities, such as Indigenous populations, principles of data sovereignty (control by the community over its own data) must be respected and supported.
Address Health Equity: When sharing health-related data, it is important to standardize and utilize information on Social and Environmental Determinants of Health (SDoH/EDoH). This data can help researchers understand and work to reduce health disparities. However, care must be taken to ensure that data sharing does not inadvertently increase the risk of re-identification for individuals in small or vulnerable communities.

Guideline: Develop a comprehensive data management plan

Develop a comprehensive data management and sharing plan before a project begins. This plan should clearly detail how data will be collected, processed, stored securely, anonymized or de-identified if necessary, shared (and with whom, under what conditions), and protected throughout its lifecycle. Ethical considerations and participant rights must be central to this plan.

How Organizations Can Support Ethical Practices

Individual data scientists need supportive organizational environments to implement ethical practices effectively. Organizations play a critical role in fostering a culture of responsibility.

Microsoft's AI for Sustainability Playbook

Microsoft's playbook offers a useful model for how organizations can approach AI responsibly, particularly in the context of sustainability, which has strong social good implications. Their five "plays" are:

Invest in AI to find solutions for sustainability.
Develop digital and data infrastructure that allows for inclusive use of AI for sustainability.
Minimize the resource use (energy, water) of AI operations and support the local communities where data centers are located.
Advance AI policy principles and governance that support sustainability goals.
Build the workforce's capacity to use AI effectively for sustainability challenges.

Google also runs an "AI for Social Good" program, stating a belief that AI can meaningfully improve people's lives. They offer various tools and support projects in areas like environmental resilience (e.g., the Heat Resilience Tool, which uses AI with satellite and aerial imagery to help cities identify ways to reduce urban heat), disaster response (e.g., Google Person Finder, which helps reconnect people after disasters), and making public data more accessible (e.g., Google Public Data Explorer).

General Organizational Practices for Ethical DSSG

Establish Ethics Review: Implement internal ethics review boards or clear processes for evaluating the ethical implications of all DSSG projects.
Provide Ongoing Training: Offer regular ethics training for data scientists, project managers, and anyone involved in DSSG work. This training should cover topics like bias, privacy, transparency, and the specific ethical frameworks relevant to their work.
Foster a Culture of Responsibility: Promote an organizational culture where ethical considerations are openly discussed, valued, and integrated into everyday decision-making.
Support Transparency Tools: Encourage and support the use of tools and practices for AI assurance (verifying that AI systems work as intended and safely) and algorithmic transparency (making AI decision-making processes understandable).

Guideline: Create an ethical environment

Organizations involved in DSSG should actively create and maintain an environment where ethical considerations are central to every project, from its initial idea through development, deployment, and ongoing monitoring. This includes providing necessary resources, comprehensive training, clear ethical guidelines, and robust oversight mechanisms.

Guideline: Providing actionable advice and steps

Frame guidelines as direct instructions or recommendations. Use clear, simple verbs to make them easy to follow. If a process is complex, break it down into manageable steps.

For example: "To protect data privacy in your project, follow these steps: 1. Collect only the data you absolutely need for your project's goals. 2. If possible, remove personal identifiers from the data (anonymize or de-identify it). 3. Store all data securely, using strong passwords, encryption, and access controls. 4. Always get clear and informed permission (consent) from people before you collect or use their data, explaining how it will be used and protected."

The field of Data Science for Social Good is dynamic, with new technologies and approaches continually emerging. Looking ahead, several trends suggest both exciting opportunities and important considerations for how data science can continue to contribute to a better world.

The pace of technological change continues to be rapid, and several emerging trends are likely to shape the future of DSSG. The Future Today Institute's 2024 report describes a "supercycle" driven by the convergence of artificial intelligence, biotechnology, and an expanding ecosystem of interconnected wearable devices. This convergence, they suggest, will redefine many aspects of life and open up new avenues for social impact initiatives.

One significant shift predicted is from the current focus on generative AI (AI that creates content) to "generative biology." In this future, AI models could design and create novel molecules, new drugs, and advanced materials. Such capabilities could revolutionize healthcare by enabling the rapid development of personalized medicines or new therapies for diseases. It could also impact material science, leading to the creation of more sustainable or effective materials for various social good applications.

Continued advancements in AI itself, including more capable multi-modal AI (systems that can process and integrate information from different sources like text, images, and audio simultaneously) and self-improving AI agents (systems that can learn and adapt on their own), will provide even more powerful tools for tackling complex social problems.

The outlook presented in a 2024 article by Interview Kickstart suggests that the importance of data science for advancing society will continue to expand. Initiatives that connect data specialists with real-world social problems are seen as a key driver of this movement, allowing data science to make a positive impact on a global scale.

Furthermore, the integration of data science with other emerging technologies, such as the Internet of Things (IoT) and blockchain, holds the promise of unlocking even greater possibilities for positive social and environmental impact. For instance, blockchain technology is already seeing applications for social good in areas like promoting sustainable tourism by ensuring fair revenue distribution to local communities, and in improving resource management through transparent and traceable supply chains.

As the field matures, there will likely be an increasing demand for DSSG projects to demonstrate tangible, equitable, and sustainable impact. It will not be enough to simply apply a novel technology to a social problem; stakeholders and funders will increasingly require rigorous evidence that these interventions are leading to real, positive, and lasting change. This means that DSSG projects will need to incorporate robust monitoring and evaluation frameworks from their inception to measure their outcomes and ensure that benefits reach the intended populations fairly and that solutions are sustainable over time.

How we can all help data science create a better world

Creating a better world with data science is a collective responsibility that involves individuals, organizations, policymakers, and funders.

For individuals:

Support Responsible Organizations: Seek out and support non-profits, social enterprises, and other organizations that use data responsibly and ethically for social good.
Practice Data Mindfulness: Be aware of your own digital footprint and how your data is being collected and used. Advocate for strong data protection and ethical data practices from companies and governments.
Contribute Your Skills: If you have data science skills, consider volunteering your time and expertise to DSSG projects or mentoring others who are interested in the field. Platforms like CitizenScience.gov or organizations listed in the Civic Tech Field Guide can offer opportunities.

For organizations (non-profits, governments, companies):

Invest in Capabilities and Training: Invest in building data science capabilities within your organization and provide ongoing ethical training for staff involved in data projects.
Foster Collaboration: Collaborate with other organizations, academic institutions, research centers, and, most importantly, with the communities you serve. Data collaboratives, where multiple organizations pool data and expertise to address shared challenges, are a growing and effective trend.
Prioritize Transparency and Accountability: Ensure that all data projects are conducted with transparency and that there are clear lines of accountability for their outcomes.
Support Open Data: Where appropriate and safe, support open data initiatives that make non-sensitive data available for public use and innovation.

For policymakers and funders:

Develop Strong Ethical Governance: Create and enforce robust ethical guidelines, regulations, and laws for the use of AI and data. These frameworks should protect human rights, promote social good, and be adaptable to technological advancements.
Fund Research and Development: Provide funding for research and development in DSSG, with a particular focus on projects that address areas of high social impact and promote the development of ethical and trustworthy AI.
Support Capacity Building and Diversity: Invest in programs that build data science capacity, especially in underserved communities and organizations. Support efforts to diversify the data science workforce to ensure a wider range of perspectives are included in solving social problems.

The journey of using data science for social good is dynamic and evolving. Many of the most pressing social and environmental challenges, such as climate change, pandemics, and human rights issues, are global in nature and transcend national borders. Similarly, data, algorithms, and AI talent flow globally. Without strong international cooperation on ethical standards, data sharing protocols, and collaborative research efforts, progress in DSSG may be fragmented, and the risks associated with powerful technologies (like the misuse of AI) could be more difficult to manage. Therefore, the future of DSSG will likely involve, and indeed require, more international partnerships, the development of shared data platforms (with appropriate privacy and security safeguards), and concerted efforts to harmonize ethical and regulatory approaches. This global collaboration is essential to ensure that data science benefits all of humanity equitably.

Even as AI systems become more sophisticated and capable of making complex decisions, the role of human agency and human oversight will remain absolutely paramount. This is critical to ensure that DSSG initiatives truly serve human values. "Social good" is ultimately a concept defined by humans, rooted in values such as justice, equity, dignity, and well-being. The consistent call from ethical frameworks for human oversight, guidance by human values, and meaningful public participation underscores the understanding that technology alone cannot, and should not, determine what is "good" for society. The future of DSSG is not about replacing human judgment with AI, but about creating powerful human-AI collaborations where technology serves as a tool to achieve human-defined goals. This requires ongoing societal dialogue about values and a commitment to ensuring that humans remain in control of how these powerful technologies are developed and deployed for social impact.

The challenge ahead is to ensure that these powerful data-driven tools and technologies are developed and deployed in ways that truly benefit all of humanity, respecting rights, promoting fairness, and contributing to a more equitable, sustainable, and prosperous world for everyone.

Conclusion

Data Science for Social Good has established itself as a vital and rapidly evolving field. Over 2023 and 2024, advancements in AI, geospatial analytics, and citizen science have opened new avenues for addressing complex societal challenges in health, human rights, education, and environmental protection. We see a clear trend towards more formalized training, a broadening scope of applications, and an increasing emphasis on community-centered approaches. The convergence of technologies like AI with other fields such as biotechnology and IoT promises even more transformative solutions.

However, the power of these tools comes with significant responsibilities. Ethical considerations surrounding bias, privacy, transparency, and accountability remain paramount. Challenges related to data quality, resource constraints, and even the environmental impact of data science itself must be proactively managed. Global ethical frameworks from organizations like UNESCO and the OECD, alongside national strategies and corporate initiatives, provide crucial guidance for navigating this complex landscape.

The future of DSSG hinges on several key developments. Firstly, there will be an increasing need to demonstrate tangible, equitable, and sustainable impact, moving beyond novel applications to proven outcomes. Secondly, global collaboration and the development of international norms will be essential, as many social and environmental challenges, as well as data and AI capabilities, transcend national borders. Finally, even as AI becomes more sophisticated, human agency, oversight, and values must remain central to ensure that technology serves humanity's best interests.

To truly harness data science for a better world, a collective effort is required. Individuals can contribute by supporting responsible organizations and advocating for ethical data practices. Organizations must invest in ethical training, foster collaboration, and prioritize transparency. Policymakers and funders play a critical role in establishing strong governance, supporting research in high-impact areas, and building a diverse and skilled data science workforce. By embracing these responsibilities, we can guide the development and application of data science to create a more just, equitable, and sustainable future for all.

Frequently Asked Questions

DSSG uses the same data analysis tools and methods as regular data science, but focuses specifically on solving societal problems like improving healthcare access, protecting human rights, addressing climate change, or promoting education equity. The key difference is the mission: instead of maximizing profits, DSSG aims to create positive social impact and help communities.

I'm interested in DSSG but don't have a formal data science background. Can I still contribute?

Absolutely! The field includes "citizen data scientists" - people who use user-friendly tools to contribute to projects. You can start by volunteering with organizations listed on platforms like CitizenScience.gov, joining community mapping projects, or participating in hackathons. Many programs also provide training, like the University of Chicago's DSSI Summer Experience for undergraduates.

What skills do I need to work in DSSG?

Core technical skills include basic statistics, data analysis (often using Python or R), and data visualization. Equally important are soft skills like understanding community needs, ethical reasoning, communication, and domain knowledge in areas like public health, education, or environmental science. Many successful DSSG practitioners combine technical skills with deep knowledge of the social issues they're addressing.

Can you give me concrete examples of how DSSG has made a real difference?

Recent examples include: Stanford students using satellite imagery and machine learning to detect illegal charcoal production linked to human trafficking in Brazil; TREND Community using AI to analyze social media conversations to better understand rare diseases like Sjögren's syndrome; and the Invisible Institute using data analysis to expose how Chicago police handle missing persons cases, leading to a Pulitzer Prize-winning investigation.

How do I know if a DSSG project is actually making a positive impact?

Look for projects that: involve the affected communities in design and evaluation, use rigorous measurement methods to track outcomes, publish their methods and results transparently, address real needs identified by communities (not just what seems interesting), and show sustainable, long-term benefits rather than just short-term fixes.

What types of organizations do DSSG work?

DSSG work happens across many types of organizations: non-profits (like TREND Community), academic institutions (Stanford DSSG, University of Chicago), government agencies (NIH, EPA, CDC), journalism organizations (Invisible Institute), tech companies (Microsoft AI for Sustainability), and social enterprises. Many projects involve partnerships between these different types of organizations.

Do I need expensive software or powerful computers to do DSSG work?

Not necessarily. Many DSSG projects use free, open-source tools like Python, R, and platforms like Google Colab (which provides free access to computing power). Cloud platforms often offer credits for social good projects. The key is starting with the problem and community needs, then finding appropriate tools, rather than requiring the most advanced technology.

Start by checking government open data portals, academic databases, and organizations working in your area of interest. If data doesn't exist, consider: partnering with organizations that collect relevant data, using surveys or community engagement to gather information, utilizing publicly available sources like satellite imagery or social media (with proper privacy protections), or advocating for better data collection systems.

Traditional methods work well for structured data and smaller datasets, while AI (especially machine learning) excels with large, complex, or unstructured data like images, text, or sensor data. For example, detecting illegal deforestation requires AI to process satellite images, while analyzing community health surveys might only need traditional statistical methods. Choose based on your data type and problem complexity.

How do I ensure my DSSG project doesn't accidentally harm the communities I'm trying to help?

Follow key principles: start by listening to and involving the community throughout the project, not just at the end; ensure your team includes people from the affected community; regularly audit your data and algorithms for bias; be transparent about your methods and limitations; have clear accountability measures; and plan for what happens if something goes wrong.

What should I do if I discover bias in my data or algorithm?

First, acknowledge the bias openly rather than trying to hide it. Then: examine your data sources and collection methods, use techniques to detect and measure bias, consider using more representative data, apply algorithmic fairness techniques, involve affected communities in developing solutions, and document your efforts to address the bias transparently.

How do I balance being transparent about my methods with protecting people's privacy?

Use techniques like: removing personal identifiers from datasets (anonymization), sharing aggregate results rather than individual data points, using privacy-preserving methods like federated learning, getting informed consent that clearly explains how data will be used, implementing strong data security measures, and being transparent about your privacy protection methods without revealing the actual data.

What career paths exist in DSSG?

Career paths include: data scientist at non-profits or social enterprises, researcher at academic institutions, policy analyst using data in government, consultant helping organizations implement DSSG projects, journalist specializing in data-driven investigations, program manager overseeing DSSG initiatives, or entrepreneur starting a social impact organization that uses data.

How do I transition from regular data science to DSSG?

Start by volunteering on DSSG projects to gain experience and understand the unique challenges. Take courses on ethics and social issues. Partner with non-profits or community organizations. Attend DSSG conferences and meetups. Consider additional training in areas like public health, environmental science, or social policy relevant to your interests.

Is DSSG a stable career field, or just a trend?

DSSG is growing into a stable field. Increasing formalization through university programs, government initiatives (like NIH's strategic plan), corporate social responsibility programs, and growing recognition of data's power to address social issues all suggest long-term stability. However, it requires continuous learning and adaptability as both technology and social challenges evolve.

How do I start my first DSSG project?

Begin by: identifying a community or organization you want to help, spending time understanding their actual needs (not what you assume they need), starting small with a focused, manageable problem, forming a diverse team that includes community members, securing necessary permissions and partnerships, and planning for how you'll measure and sustain impact.

What are the most common reasons DSSG projects fail?

Common failure reasons include: not involving the affected community from the beginning, focusing on technical solutions without understanding the real problem, poor data quality or inappropriate data, lack of resources for implementation and maintenance, ethical issues that erode trust, and failure to plan for long-term sustainability after the initial project ends.

How do I measure the success of a DSSG project?

Define success metrics early that include: quantifiable social outcomes (lives improved, problems solved), community satisfaction and engagement, sustainability of the solution, ethical compliance, and lessons learned. Use both quantitative measures (numbers) and qualitative feedback (stories, interviews). Plan for long-term follow-up, not just immediate results.

What new technologies should I learn about for future DSSG work?

Emerging areas include: explainable AI (XAI) for transparency, federated learning for privacy-preserving collaboration, geospatial analytics for location-based insights, generative AI for creating educational content or communications, multimodal AI that combines different types of data, and edge computing for deploying AI in resource-limited settings.

How is the field of DSSG expected to change in the next 5-10 years?

Expected changes include: greater emphasis on proving tangible impact rather than just applying new technology, more international collaboration on global challenges, stronger ethical frameworks and regulations, integration with emerging technologies like IoT and blockchain, increased focus on environmental sustainability of data science itself, and more formalized career paths and educational programs.

Both approaches work. Specializing allows you to develop deep domain expertise and relationships within a specific area like public health or climate change. Staying broad lets you apply your skills across different problems and learn from various contexts. Consider your interests, career goals, and local opportunities. You can always start broad and specialize later, or vice versa.

How do international ethical frameworks like UNESCO's affect DSSG work?

These frameworks provide important guidance for responsible practice, especially for projects that cross borders or involve sensitive data. Key principles include respecting human rights, ensuring fairness and non-discrimination, protecting privacy, maintaining transparency, and requiring human oversight. Familiarize yourself with relevant frameworks and use them as checklists for your projects.

How can DSSG address global challenges like climate change that cross national boundaries?

Global challenges require international collaboration through: shared data platforms (with appropriate privacy protections), harmonized ethical standards across countries, collaborative research partnerships, common measurement frameworks, and coordinated policy responses. Organizations like the NIH and initiatives like Microsoft's AI for Sustainability demonstrate how this collaboration can work in practice.

Harnessing Data to Build a Better World: A 2024 Look at Data Science for Social Good

Share this article