A Q&A with a UCIPT Postdoctoral Researcher
Is there a correlation between Twitter sentiment and athletic performance?
I am not aware of any research studies that have examined this question, but I have witnessed a trend between Twitter activity and athletic performance. When a particular sport team is doing well, there is usually a spike in tweets/hashtags related to that team and the sentiment is usually positive. For example, when the Toronto Raptors (I am huge Raptors fan since I am from Toronto) won in the first round of the NBA playoffs against the Indiana Pacers, I saw a huge spike in the hashtags #WeTheNorth and #goraptorsgo. However, it is currently challenging to predict athletic performance using tweets alone. I think tweet data could be extremely valuable if it were combined with other sports-related data (e.g., historical team performance, current team injuries).
What advice would you give UCLA’s athletic directors when trying to decipher the Twitter feeds of competitor teams’ players?
Twitter can be a great tool to gain insights into people’s attitude and behavior. One of our recent studies have shown that the types of tweets can be used to measure stress levels. Thus, I think it is possible to analyze Twitter feeds of a team’s players to understand their stress level and emotional state (e.g., anger, sadness, happiness, fear). Obviously, this would be helpful for an athletic director looking to understand how the players are feeling. Twitter can also be a great tool for understanding public opinion about a team. I think UCLA’s athletic directors could use this to motivate players as well.
What Twitter analysis tools are used by professional gamblers when reviewing player feeds? Is there something that academic researchers can learn from data analysts who are motivated by real-time financial risk?
I am currently unaware of any Twitter analysis tools that are used by professional gamblers when reviewing player feeds. However, a recent study reported that Twitter activity (sentiment analysis) was able to predict the general trends of the financial market. Thus, I can see that a real-time analysis of sentiment might be useful for professional gamblers. However, I think more research is needed to establish whether Twitter analysis tools can be useful for gamblers.
Do Twitter sentiment analysis tools take longitudinal information, so that it is possible to see the frequency and timing of tweets and estimate if an athlete is getting enough sleep? What other physiological information might be culled out of Twitter data?
I think it is possible to determine whether an athlete is getting enough sleep using longitudinal data from Twitter. Each tweet is time stamped, thus if a player is tweeting at 3:00 A.M., this may give an indication that the athlete is not getting enough sleep. However, there are limitations to this method of measurement since it is not a direct measurement of sleep quality. Therefore, I think devices such as wearable watches give a better indication of sleep. Many of these wearable devices enable users to share their results on social media. This feature offers the potential to measure someone’s sleep patterns.
What tools or methods might exist to map physical or fitness progress of an athlete using Twitter and/or Twitter sentiment analysis? Are there tools that are already in use?
There has been a recent study that used Twitter to measure physical activity and obesity levels across the United States at the county level. The authors of the study reported that a greater number of physical activity–related tweets within a county was related to lower rates of obesity. Currently, there are no tools that map physical progress at the individual level using Twitter and/or Twitter sentiment analysis. As mentioned earlier, I think in order to track fitness progress of an athlete, we will most likely need to combine Twitter data with data from other devices that measure physical or fitness progress directly (e.g. heart rate, maximum oxygen consumption, lactate threshold).