Facebook Details New Data for Good Initiatives Related to Covid-19

Facebook released new visualizations and datasets that are part of its Data for Good initiative, as well as a new survey related to Covid-19. The social network said in an updated Newsroom post that its new Covid-19 Map and Dashboard incorporates international results from the symptom survey conducted by Carnegie Mellon University Delphi Research Center in April, as well as its datasets on movement range that are helping inform public-sector response to the pandemic. More mobility datasets were released to show the rates at which different communities are reducing their mobility or remaining in the same place. Facebook said these datasets use aggregated data, and it applied a differential privacy framework to protect users’ privacy when the datasets are created and shared. [kickout id="1"] Facebook released a new map displaying travel patterns between countries and states to help researchers and non-governmental organizations understand the impact of long-distance travel on the spread of the coronavirus. The social network shared new insights from partners that have been using its disease prevention maps to analyze how Covid-19 may continue to spread and its effects on communities. Finally, Facebook is teaming up with the Initiative on the Digital Economy at the Massachusetts Institute of Technology on a new survey about people’s knowledge, attitudes and practices regarding Covid-19, with advisement provided by the Johns Hopkins University Center for Communication Programs and the World Health Organization. Data for Good policy lead Laura McGorman addressed the program’s privacy aspects in a separate Newsroom post. McGorman said protected datasets are only shared with Facebook’s network of trusted research partners, and those partners can only access aggregate information, not individual information. She added that some datasets are shared publicly, but they are formatted in order to prevent reidentification while preserving the insights needed to help respond to crises. McGorman wrote, “For example, datasets that include location information are aggregated in a way that protects the privacy of individuals by using techniques like spatial smoothing to create weighted averages and avoid using map tiles where very few people live. And for public datasets on mobility, we use our differential privacy framework, which takes into account the sensitivity of the aggregated dataset and adds a random number of additional observations to each map area in order to ensure that no one can reidentify users.” The location data used in Data for Good maps is the same as the data Facebook uses to show locally relevant content to users, and people can opt out via the location history setting. McGorman said data collected for Data for Good initiatives is restricted to information that can help researchers, academics and other trusted partners address humanitarian crises and social issues, adding that no additional data is collected and none of it is used to impact advertising. She wrote, “Some data used in Data for Good is the same data that we use to personalize your experience on our applications and show you more relevant content and ads. For example, if you choose to share your location information, that information may be used in a Data for Good map or dataset, but its inclusion in Data for Good does not impact the ads or content you see on our apps.”