We were honored to join more than 500 passionate folks from a variety of industries at Bloomberg’s 5th annual Data For Good Exchange Conference. Free and open to the public, the event is devoted to solving social problems through the use of data science.
This year’s theme was “Our Data for Good?”, focusing on the ethical and equitable use of data to benefit society. Presentations, papers, and interactive workshops covered an array issues, such as socioeconomic status, gender, race, and immigration. We enjoyed the conference’s spirit of collaboration, the sharing of ideas, and a general feeling that ‘we’re all in this together’.
Rose Gill Hearn, principal of Municipal Integrity at Bloomberg Associates, kicked things off with her keynote address about cities using data for good. She spoke about the impact that the Data for Good Exchange programming has had on the global community. According to Hearn, though cities generate a plethora of data, that data tends to go underutilized. She explained how Data for Good has helped cities solve significant data challenges to inform policy, create more transparency, and improve delivery of services. As an example, Hearn cited a partnership with Rio de Janeiro aimed at improving their business licensing process, transitioning them from a complex and paper-based system to a streamlined online portal.
Following Hearn’s remarks, we heard from an all-women panel of representatives from MasterCard, the World Bank, and New America’s Public Interest Technology. They discussed how data can be used to address economic shocks and support resilience within communities. Nina Klapper, Lead Economist at the World Bank, described the ways in which mobile finance supports the financial independence of women in developing countries. Shamina Singh, President of the Center for Inclusive Growth, talked about the use of transaction data to prevent price gouging in areas hit by natural disasters. Michelle Thompson, Public Interest Technology Fellow at New America, highlighted the importance of local perspective in any data effort. For example, if data collectors working in disaster recovery are unaware that some regions don’t use deeds, but pass property down from generation to generation without formal records, how can those communities be meaningfully served by modern data methods?
Also of note was the discussion of the Equity Intelligence Platform, a joint effort by the Obama Foundation and Brighthive. Panelists shared how data from cities can be incorporated into proactive measures that improve outcomes for young men and boys of color. “The data does not belong to the city. It belongs to the people,” said Niiobli Armah IV of Bloomberg Associates. Cyrus Garrett, manager of My Brother’s Keeper Alliance at the Obama Foundation, said that while government have broken down some silos in their data, social services still have that issue. Garrett highlighted how the Obama Foundation, “…as convener and thought partner, helps identify how to ethically use data so we can enter in these conversations with communities—not to do it for them but do it with them.”
Even with all of the new tools and data sources available, it’s still paramount that analysts ask the right questions. In a panel about sustainable finance, Bloomberg Quantitative Researcher Arun Verma demonstrated pitfalls of data science by noting spurious correlations like ‘per-capita cheese consumption’ and ‘number of people who died by becoming tangled in their bedsheets’.
Bloomberg also debuted their 2018 Gender-Equality Index. It recognizes 104 companies who exhibit exemplary workplace gender equality based on factors like paid paternity leave, healthcare services, and gender reassignment policies. Leveraging their role in the financial services industry, they make these metrics available to traders on the stock market floor and beyond. A previous iteration of the index showed improvement across these equality metrics from 2014-2016.
For the first time, the conference included workshops to give attendees opportunities to interact with methodologies, principles, and practices. Since it was impossible to attend all of the workshops (some of which happened concurrently) a panel was dedicated to discussing the main ideas and takeaways from each. Some of the key ideas included:
Data, algorithms and people are always biased. Data is manipulated and analyzed by people. To be responsible practitioners of data, we must recognize the inevitable biases in our work.
Diversify the data community. Bring together a heterogeneous group of people who can speak from diverse perspectives. Data scientists may be the ones working directly with the data, but citizens, advocates, and practitioners have a stake in the outcome of the analysis. If data is truly to be used for good, that process must include everyone.
Bring data literacy to communities. This is a key pathway to trust. By improving data literacy, you’re building a bridge between the data and the folks who are reading/receiving that data. If you want more people to be involved with the data, you have to make sure they can understand it.
Address trust issues between government and citizens. As Jake Porway of DataKind said, “Whatever you do for me, without me, you do to me.” Encourage more transparency and interaction between citizens and data. Counter misinformation and disinformation by showing people why their data is being collected and how it serves the greater good.
Karin Klien, founding partner of the Bloomberg Beta venture firm and a keynote speaker, spoke about the opportunities presented by cross-sector collaboration. Together, nonprofits and government agencies who have insight can be matched with data scientists, academics, and practitioners who have the tools to analyze their data. Working together, they can design and deploy data tools that benefit communities and citizens in a lasting way.