UMD Data Challenge 2021 - COVID-19 Global Symptoms Tracker Team42:
- Gabriel Sestieri
- Theodore Gaidis
- Manar Al-badarneh
- Brendan Goodhue
University of Maryland’s Joint Program in Survey Methodology and Carnegie Mellon University’s Delphi Research Group, collaborated with Facebook to invite people to participate in surveys that ask about how they are feeling, including any symptoms they or members of their household have experienced and their risk factors for contracting COVID-19. The surveys are designed to provide valuable information to help monitor and forecast how COVID-19 may be spreading, without compromising the privacy of the people who took the surveys. Facebook does not share who took the surveys with our academic partners, and the universities do not share individual survey responses with Facebook. With over 2 billion people on Facebook, Data Challenge Competitors are in a unique position to use the data to study COVID-19, such as how the pandemic is affecting population movement trends, and understand which areas may be at risk of an outbreak based on population characteristics and symptoms.
Our research aims to identify and analyse the relationships between COVID-19 indicators from the UMD COVID-19 World Survey Data API to inform the audience about certain patterns and trends we found. The COVID-19 World Survey Data API contains 21 indicators recorded from a Facebook survey answered from around the world. We automated a script to consolidate the data for each country through iterative API calls. Each CSV file contains 23 columns including fields such as country name, date of each reporting, and the smoothed data for each of the 21 indicators on each date. We will cross compare the indicator values between countries to find insights into how their response to the pandemic affected these indicators, and how these indicators may uniquely impact one another in each country we analyze. This format of data also allows us to perform time series analysis. Using data visualization tools will allow us the ability to express our findings of our technical work for the general public to easily understand and interpret. Our analysis can help to prepare for a rise or fall in people experiencing symptoms and ultimately help policymakers and Public health officials in determining which country or region would benefit the most from added healthcare assistance.
For the 2021 UMD Data Challenge our team chose to analyze the level three, COVID-19 World Survey Data API. In order to acquire data for this project we built a data scraper. It used a series of loops to build the links necessary to call the API for every data point it contained. The program built links for each of the 21 indicators for every date and country available. After cleaning the data and getting it into workable CSV formatting we began to narrow down our problem statement. Our motivation to solve these problems came from wanting to make an impact in the fight against COVID. People all over the world are dealing with the impact and burden of disease that COVID-19 has left. Using the data available to us we were motivated to help the general public and Public Health officials with our data analysis and findings.
Additional data analysis led us to focus on three primary questions for our problem statement; the first being, for countries in the Schengen Area (the countries in Europe that have open borders with one another), does a spike in COVID-19 related indicators in one country correlate to a delayed spike in indicators in their neighboring countries? The next question we posed was do the people’s trust in government, trust in healthcare officials, and trust in the WHO, have an effect on the number of people willing to take a vaccine? To tie together the rest of our findings we asked how do social behaviors (contact with someone outside your household) correlate to mask wearing habits, COVID-19 cases in the community, and financial worries?
In analysis of our first primary question we came to the conclusion that there is a correlation in COVID-19 spikes between neighboring countries in the Schengen region. All countries of the region had similar trends in the indicators we deemed most important to focus on. Next we strove to answer how notable figures and peoples' opinions of trust in the vaccine affect their peers' willingness to take the vaccine; we found healthcare officials garnered the mosttrust while politicians, the least. Finally, in our analysis of survey respondents mask wearing habits we found that in summer months mask wearing decreased however this did not lead to a notable increase in COVID-19 like symptoms.