by Riya Sirdeshmukh
In today’s world, data is power, making organizations—private and public—rely heavily on a data-driven decision-making approach, which involves collecting, extracting patterns and insights from data, and using those insights to inform decision-making. Without data, organizations are more likely to make decisions based on intuition and gut feeling that allow bias and noise to hamper organizational growth. A more equitable approach would be to let intuition guide while using data to verify, quantify, and understand related objectives. Data-driven decision-making is effective when trying to forecast trends in industries, assess customer responses before rolling out a new product, find and rectify sources of systematic error, and, importantly, achieve social justice.
Social justice is the concept of fair, and equal relations between people measured by the distribution of wealth, opportunities for advancement, access to healthcare, and other privileges within a society. The social justice movement, encompassing the Black Lives Matter Movement and the #MeToo Movement, among others, is a fight to break down barriers that prevent equal opportunity based on factors such as race, economic status, gender, and sexual identity, making data a powerful weapon in this fight. Although individual stories on similar issues are making the rounds with its increasing numbers of data. Data, when documented, analyzed, and reported effectively, reveals the full scope of the issue, thus driving concrete and measurable change across boards.
_________________________________________________________________________________________________________
Using Data to Support the Black Lives Matter Movement
Many organizations have made efforts to use data to help the Black Lives Matter Movement. To highlight a few:
To understand how and why racism manifests on Airbnb’s platform, Project Lighthouse launched a data analysis project on June 30, 2020 that aims to study customer usage, find any patterns and trends relating to race. The insights gotten from this will inform policy and design to address any racial inequity that emerges.
Data Science for Good: Center for Policing Equity
In 2018, via Kaggle, an online data science platform owned by Google, the Center for Policing Equity (CPE) incentivized the data science community to solve the problem of racism within policing by providing data as well as $15,000 for the most innovative solutions. The data CPE provided contained documented field interviews with information on educational attainment, occupation, income, housing status, and level of poverty. Data scientists had to attempt to pinpoint the areas and instances of racial disparities in policing. The project was successful in informing police agencies where they could make improvements by identifying deployment areas where racial disparities existed.
Justice of Policing Act of 2020
The main view described under the act is to increase police transparency through the collection of data. Furthermore, one of the directions in which the state of Boston has reallocated police funding is towards, “analyzing data to better understand the inter-connectedness of societal, environmental, and behavioral factors that contribute to the impact of racism, and access to jobs, food, housing, transit, and education.” In these instances, data analysis is an important vehicle for police reform.
With the knowledge that great power comes with great responsibility, makes these examples exhibit the immense power that data has to tackle global issues.
Data in the Wrong Hands Oppresses
Data is a double-edged sword. Without it, organizations can not invest, streamline, and improve products or effectively reach their audience. On the other hand, data also has the power to divide, marginalize, exclude, and oppress. For example, bias can creep into data at any point in the data-to-insight lifecycle, and secondly, big data insights can be misused to harm specific populations.
When using data to solve a social issue, the issue must first be accurately defined. Then, data is collected, cleaned, explored, and prepped for learning. Subsequently, the prepped data is modeled and visualized, and lastly, analyzed for insight. Although attention has been brought to problems with machine learning and artificial intelligence (AI) wherein feeding skewed data to a computer can produce biased algorithms. With every massive data come human involvement, and the bias nature of humans can creep in at any stage of the process. More specifically, the definition of the issue, the data points collected, the analysis, and the factors that are omitted while cleaning data are all impacted by the biases of the person performing these actions.
As quoted by Jamelle Watson-Daniels, Director of Research at Data for Black Lives, “we should not treat data as if it has been collected in a vacuum separate from society nor should we assume that algorithms are inherently fair or equitable because they are not human.” When data scientists are not representative of the world’s diverse population (a.k.a. rich, white, cis, heteronormative male data scientists), data is bound to be twisted and insights gathered will oppress.
Even if one manages to form a group of data scientists who create a diverse dataset, data misuse is another problem. After huge amounts of data about human behavior are mined, a wealth of information and insights can be collected from this data. When this information falls into the wrong hands, big data can be misused and misinterpreted. The famous Cambridge Analytica Scandal is a poster child for data misuse and weaponization of data. Cambridge Analytica purchased Facebook data on millions of Americans (without their knowledge) and used it to target voters who pushed votes for Ex-President Donald Trump.
Organizations Aiding Data-Driven Decision Making for Social Change
In response to George Floyd’s murder which fueled the acceleration of the Black Lives Matter movement, many organizations are working tirelessly at the intersection between data science and social justice with the aim of reducing bias, misuse, and overcoming hurdles to effectively harness data. This could further propel this movement to support and uplift Black lives.
Data for Black Lives, founded by Yeshimabeit Milner, in 2015, is a growing movement focused on giving Black Americans a seat at the table for all types of data-related activities, from collection to interpretation to application. They are doing path-breaking work to tackle the issues relating to data and Black lives from multiple fronts. Additionally, many other organizations are supporting Black lives from various angles. To highlight a few problems and their solutions:
Problem: Before data can be used for social change, it must be collated. A big challenge is that the data is not all in the same place. The effort takes to find, collect, and curate that data is sometimes far more than a small social justice organization can give.
Organization: Institute for Development Impact (I4DI) situated at the nexus of research and technology is one of many organizations working hard to collate data relating to the Black Lives Matter Movement. Specifically, the Data for Social Change Project aims to develop a historical archive/database to document, map, and visualize the tangible results of the global protests and racial justice activism in the wake of the murder of George Floyd.
Problem: Data science has a major inclusion problem. Under-representation of marginalized people hampers the process and twists results
Organization: Correlation One introduces a new data science program, Data Science for All / Empowerment, providing Black, Latinx, and LGBT+ individuals the skills and connections to start and advance their careers in the data science and analytics fields.
Problem: Humans are biased and hence, data is biased.
Organization: The Columbia University Data Science Institute recognizes the biased nature of data and emphasizes collaboration across disciplines to shift the exact questions asked and interpretation of results.
From collecting new insights to reframing conversations and asking tough questions, there is immense strength in numbers as seen in Black Lives Matter. Now is the time to use data to drive change.