Interview Questions for Cornelius Fritz, Statistician at the University of Munich Regarding Facebook's Data for Good Initiative

In a groundbreaking study published in January, researcher Cornelius Fritz utilised data from Facebook to analyse the impact of regional differences in mobility and social connectedness on the spread of COVID-19 in Germany.

The study, which incorporated three types of data - colocation maps, social connectedness index, and geolocation data - aimed to identify areas where people were staying put. The data was aggregated to the 401 federal administrative districts in Germany, allowing for a detailed analysis of the country's regions.

One of the study's most intriguing findings was a distinction between East and West Germany in terms of social ties, even more than 30 years after reunification. The study suggests that the number of infections in a district would likely be lower than the national average if that district was located in the former East Germany.

To measure the strength of friendship ties between the districts of Germany, the study included a social connectedness index. Colocation maps were also constructed to indicate the probability that two people from different districts would meet up during a given week. The study found that the districts in the east of Germany had fewer connections with other districts compared to those in the west.

The study's model, which was built to forecast COVID-19 infection rates one week into the future, showed consistently smaller errors than benchmark models. This suggests that machine learning models such as graph neural networks are effective for COVID-19 forecasting, particularly when using nuanced data on behavioural patterns.

The graph constructed in the study shows social ties among western districts on one side and social ties among eastern districts on the other, with Berlin as the central city connecting the two sides. This finding could be useful for policymakers in evaluating district-level policies, particularly in terms of limiting trans-district movements and concentrating meeting patterns through local lockdowns to mitigate further national outbreaks.

Policymakers can also use this type of model as a predictive tool to better manage healthcare resources such as hospital beds, respirators, and vaccines. However, it is important to note that epidemiological data such as infection rates, recovery rates, mortality rates, demographic information, and possibly real-time health system data are necessary to use such models for COVID-19 predictions. Additionally, data about virus variants, vaccination rates, mobility, and social behaviour may be required to improve model accuracy.

Facebook's data was crucial for the study due to its granular level, which is not readily available from other companies. The data was used to identify the percentage of people who were staying within a 0.6km radius throughout a day, providing valuable insights into people's behaviour during the pandemic.

In conclusion, the study by Cornelius Fritz offers a unique perspective on the impact of regional differences in mobility and social connectedness on the spread of COVID-19 in Germany. The findings could prove invaluable for policymakers as they navigate the ongoing pandemic and work towards effective strategies for disease control and resource management.

Interview Questions for Cornelius Fritz, Statistician at the University of Munich Regarding Facebook's Data for Good Initiative