Human Movement Patterns of Different Racial-Ethnic and Economic Groups in U.S. Top 50 Cities: What Can Social Media Tell Us About Segregation?

Abstract

Human movement patterns, one of the important fields in human mobility study, is significant for various practical applications, and many studies have proven that it is strongly impacted by individual socioeconomic and demographic background. On the other hand, social media has become more and more popular in studying human movement patterns because of its exclusive advantages compared with traditional data source (e.g., travel diary and American Community Survey from U.S. Census). While many efforts have been made on exploring the influences of age and gender on movement patterns using social media, this study aims to analyze and compare the movement patterns among different racial-ethnic and economic groups using social media (i.e., geotagged tweets) from the U.S. top 50 populated cities. Since individual racial-ethnic information is usually not revealed by the social media users, we adopt different name-ethnicity prediction models to infer the race-ethnicity for the users. As for the individual economic status inference, the median house value from U.S. Census is utilized as a reliable reference. Results show that the average tweeting density (i.e., tweeting frequency divided by the population of the group) of rich groups is 17% lower than the one from poor groups, but for different racial-ethnic groups with the same tweeting density, our results reveal that Non-Hispanic Black or African American have 5% more activity zones than Non-Hispanic Two or More Races and Hispanic or Latino origin. As for median travel distance, poor groups travel 42% shorter than rich groups. On the other hand, the median travel distance of Non-Hispanic White is 23% longer than the one of Hispanic or Latino origin, 18% longer than the one of Non-Hispanic Two or More Races, 10% longer than the one of Non-Hispanic Asian and Native Hawaiian and Other Pacific Islander, 9% longer than the one of Non-Hispanic American Indian and Alaska Native. To explore further on outbound-city travels (i.e., the travels with origins inside the boundary of the Urban Areas defined by the U.S. Census but with destinations outside of that boundary), poor groups contribute 10% less outbound-city travels than rich groups. Particularly, the poor groups from the racial-ethnic minorities such as Non-Hispanic American Indian and Alaska Native (18%), Non-Hispanic Asian and Native Hawaiian and Other Pacific Islander (16%), and Hispanic or Latino origin (14%) have much lower percentages, while the poor groups from Non-Hispanic White and Non-Hispanic Black or African American reach the overall mean percentage (25%). This finding strongly proves that people who are economically disadvantaged and racial-ethnic minorities are more restricted in long distance travels, which indicates their spatial mobility is more limited into the local scale. Another important finding is the economically-segregated movement pattern in the national scale – rich neighborhoods are mostly visited by the rich, while poor neighborhoods are mainly accessed by the poor, but some race-ethnicity groups can diversify this segregated pattern in the local scale, such as the Non-Hispanic Asian and Native Hawaiian and Other Pacific Islander + Poor group in New York having a much higher percentage (38%) of traveling to rich community than the national average level of p-to-r travels (12%). Lastly, spatial variability of travel distances is also revealed. Although there is a uniform pattern of travel distance distributions among the U.S. top 50 populated cities, which are fitting a decreasing curve with long-tail, yet the median travel distance for the top 6 cities are significantly different (e.g., New York City 6245 m; Los Angeles 7362 m; Chicago 6807 m; Houston 9729 m; Philadelphia 7233 m; and Phoenix 8827 m). On the other hand, for the percentage of outbound-city travels of the U.S. top 6 cities, it shows that New York (27.2%) and Houston (27.5%) have more outbound-city travels, which could indicate their lightly stronger interaction power with other cities, while Los Angeles (22.8%) and Chicago (21.7%) have less outbound-city travels.