Understanding Human Mobility from Twitter.
Bottom Line: Our analysis shows that geotagged tweets can capture rich features of human mobility, such as the diversity of movement orbits among individuals and of movements within and between cities.We also find that short- and long-distance movers both spend most of their time in large metropolitan areas, in contrast with intermediate-distance movers' movements, reflecting the impact of different modes of travel.Our study provides solid evidence that Twitter can indeed be a useful proxy for tracking and predicting human movement.
Affiliation: CSIRO, Brisbane, Australia.
Understanding human mobility is crucial for a broad range of applications from disease prediction to communication networks. Most efforts on studying human mobility have so far used private and low resolution data, such as call data records. Here, we propose Twitter as a proxy for human mobility, as it relies on publicly available data and provides high resolution positioning when users opt to geotag their tweets with their current location. We analyse a Twitter dataset with more than six million geotagged tweets posted in Australia, and we demonstrate that Twitter can be a reliable source for studying human mobility patterns. Our analysis shows that geotagged tweets can capture rich features of human mobility, such as the diversity of movement orbits among individuals and of movements within and between cities. We also find that short- and long-distance movers both spend most of their time in large metropolitan areas, in contrast with intermediate-distance movers' movements, reflecting the impact of different modes of travel. Our study provides solid evidence that Twitter can indeed be a useful proxy for tracking and predicting human movement.
No MeSH data available.
Related in: MedlinePlus
Mentions: Longer distance movers in Fig 4(c) and 4(d) appear to have a more stretched component in the negative x-axis, with an expanding gap close to the origin as rg increases. To shed further light on this effect, Fig 6 compares the spatial distribution of tweets for the four categories of rg, focusing on the southeast region that includes more than half the country’s population (see Supporting Information S1 Text for full maps of Australia). It shows that for smaller rg, tweets are concentrated in clusters mainly in large cities, or other regional areas. For intermediate rg, we observe a much stronger tendency of tweets to be within an expanded region around key cities and along main roads connecting large cities. The tight coupling of movement at these distances with road usage explains the increased directivity (and anisotropy) of motion for these rg categories. The tweet distribution for large rg > 500km shows a completely different trend, with a renewed focus of tweet activity in and closely around the main cities. This difference likely stems from the change in mode of travel to airplanes, where people fly in to a destination with the intention of remaining within a limited orbit around this destination. These long distance movers appear to have a few target destinations, such as airports at key cities or locations of interest. As we average the movement patterns over a population, the dominant movement distances for each individual may vary widely, contributing the stretch of this tail along the negative x-axis in Fig 4(c) and 4(d). The gap that appears on the negative x-axis close to the origin arises from the use of air travel, where people do not tweet between source and destination as they travel long-distances. These patterns may also be related to the sparse and concentrated population in Australia with heavy concentration of people (and likely Twitter users ) at major population centres. Another likely cause of this pattern is the fly-in/fly-out worker phenomenon , where workers in the mining sector stay at remote sites during weekdays and then return home or travel to holiday destinations in Southeast Asia during weekends. It is also likely that the increased isotropy for distances from 200km up to 1000km is due to long distance movers circulating in the vicinity of their destination away from home, given the high cost  associated with returning to their home location. These long distance travellers may be sending tweet messages mostly from their main destination (such as the arrival airport or a remote work site), and less frequently tweeting from satellite locations.
No MeSH data available.