Commuter Mobility Patterns in Social Media: Correlating Twitter and LODES Data

Abstract

The Longitudinal Employer-Household Dynamics Origin-Destination Employment Statistics (LODES) are an important city planning resource in the USA. However, curating these statistics is resource-intensive, and their accuracy deteriorates when changes in population and urban structures lead to shifts in commuter patterns. Our study area is the San Francisco Bay area, and it has seen rapid population growth over the past years, which makes frequent updates to LODES or the availability of an appropriate substitute desirable. In this paper, we derive mobility flows from a set of over 40 million georeferenced tweets of the study area and compare them with LODES data. These tweets are publicly available and offer fine spatial and temporal resolution. Based on an exploratory analysis of the Twitter data, we pose research questions addressing different aspects of the integration of LODES and Twitter data. Furthermore, we develop methods for their comparative analysis on different spatial scales: at the county, census tract, census block, and individual street segment level. We thereby show that Twitter data can be used to approximate LODES on the county level and on the street segment level, but it also contains information about non-commuting-related regular travel. Leveraging Twitter’s high temporal resolution, we also show how factors like rush hour times and weekends impact mobility. We discuss the merits and shortcomings of the different methods for use in urban planning and close with directions for future research avenues.

Commuting Analysis of the Budapest Metropolitan Area Using Mobile Network Data

“…The fraction of workers who have their homes in the same district was very close to that of the census data in the outer districts (15)(16)(17)(18)(19)(20)(21)(22)(23) but generally overestimated in the core districts (1,(5)(6)(7)(8)(9) and the inner districts (2)(3)(4)(10)(11)(12)(13)(14). The workers from other district groups showed the best match to census data (where the CDR should have the best quality), while the agglomeration was somewhat overestimated in many districts.…”

“…Since these locations fundamentally determine people's mobility customs, the commuting trends can be analyzed between these locations. Commuting is studied using mobile network data within a city [14,15] or between cities [16][17][18][19] and also examined by social network data, such as Twitter [20][21][22].…”

Social media mining under the COVID-19 context: Progress, challenges, and opportunities

“…These APIs return certain percentages of their total content, with some of them containing geo-information at various levels. Studies have found that the Twitter-derived mobility patterns can approximate commuting patterns ( Petutschnig et al, 2021 ) as well as mobility records released by Apple, Google, and Descartes Labs ( Huang et al, 2021 ). Using 580 million geotagged tweets collected worldwide, Huang et al (2020) measured human mobility by proposing the concept of single-day distance and cross-day distance, which highlight the users’ daily travel behavior and the users’ displacement between two consecutive days, respectively.…”

Post-Pandemic San Francisco CBD: A Spatiotemporal Analysis on Pedestrian and Consumer Patterns

“…An effective strategy to minimize biases and errors in location-based mobile data is to integrate it with other data sources (Li et al, 2023;McMillan, 2016;Offenhuber, 2014;Petutschnig et al, 2021). In this research, consumer spending data is also obtained to supplement foot traffic data.…”

“…Literature Review Human mobility refers to the spatial and temporal movement of individuals, influenced by geographical factors such as population size, urban structure, land use, and regional connections. Its information is commonly found in various data categories, such as social media data (Bernabeu-Bautista et al, 2021;Ebrahimpour et al, 2020;Cheng et al, 2011;Terroso-Saenz et al, 2022), call detail records (Kang et al, 2010), economic statistics (Petutschnig et al, 2021), and transit trajectory data (Hasan et al, 2013). Human mobility data plays a pivotal role in various applications.…”