Understanding human mobility using Twitter

Abstract

Advances in communication technology have enabled researchers to look deeply into the dynamics of human mobility. Especially, the advent of online social media has given researchers unprecedented access to a large amount of human movements data both in space and time. By analysing this data, a better understanding of the impact of human movement can be achieved in areas such as disaster management, transportation planning, communication networks and epidemic modelling and prediction. Most of the prior studies make use of data which is either coarse-grained, proprietary or limited in size. This thesis, however, leverages a large set of publicly available location data from geo-tagged tweets in an attempt to gain deeper insights into human mobility dynamics. System-wide mobility analysis of Twitter users is conducted at the suburban level in Australian cities. Specifically, movement models such as Gravity and Radiation are employed to test the movement behaviour of users and it is found that the classic Gravity model captures user movements better than the competitor model. This is followed by individual-based mobility analysis of users by using measures such as Entropy and Predictability which signify the theoretical degree of randomness in movements. It is shown that users can be distinguished on the basis of their movement patterns and two distinct groups 'frequent' and 'casual' are identified. The prediction problem is also explored here with results suggesting that users in Brisbane have higher prediction accuracy than corresponding users in Sydney and Melbourne indicating heterogeneity among various cities. The top 10% of Brisbane's most predictable users have 76.6% prediction accuracy compared to 49.2% for Melbourne and 54.8% for Sydney. The prediction information could be useful for such users and they can be apprised of any update about their likely next location.

Pakistan in 2023: On the Brink of a Full-Blown Crisis

“…Pakistan's army chief stated that Pakistan wants an enduring relationship with the USA while meeting with a US diplomat (Reporter, 2021). Recently, TV anchors and policy analysts considered close to the military establishment have openly criticised the harsh conditions in CPEC agreements and the economic viability of the CPEC projects (Khan, 2021). These TV analysts are debating whether Pakistan should recalibrate its ties with China vis-à-vis USA, given Pakistan's massive dependence on US-led multi-lateral donor and regulatory agencies.…”

The Affective Economies of Sovereignty: Desire and Identification

“…Khan's ensuing campaign is centered around the sovereignty of Pakistan in the form of the 'Haqiqi Azadi' (translation: real freedom). 28 Shireen Mizari, a former academic, journalist, and ex cabinet member for Khan styles herself as a 'believer in sovereignty of Pakistan' as her profile on Twitter. 29 Khan describes Pakistan as a slave nation, and his fight is for its real freedom.…”

Efficient Surveillance and Temporal Calibration of Disease Response

“…The Twitter data supplements the IVS and NVS datasets as a source of intra-region mobility. It was collected in 2015 and contains 925,945 trips from 79,271 individuals [20]. For the missing years, we apply an auto-regressive moving average model [21], a common method for forecasting time-series to estimate movement between regions.…”