Barriers of Urban Mobility

Gergő Pintér¹ and Balázs Lengyel^1,2,3

¹ANETI Lab, Corvinus University of Budapest
²ANETI Lab, HUN-REN Center for Economic and Regional Sciences
³Institute for Data Analytics and Information Systems, Corvinus University of Budapest

5 September 2024
pintergreg.github.io/ccs24

all roads lead to Rome

motivation

Nagykőrösi road, Budapest by vst via Mapillary CC BY-SA 4.0 — Nagykőrösi road, Budapest
by vst via Mapillary CC BY-SA 4.0

amenities enter the equation

complex amenities foster social mixing (Juhász et al., 2023)
complex as in economic complexity
still working combining the two branches of research

motivation

A wildlife overpass built over Highway 38 in Israel by Hagai Agmon-Snir | CC BY-SA 4.0 — A wildlife overpass built over Highway 38 in Israel
by Hagai Agmon-Snir | CC BY-SA 4.0

mobile positioning data

collected from various, unspecified smartphone apps
- timestamp, user ID, location
- GPS-based location
pings are clustered into stops (Juhász et al., 2023)
- using Infostop algorithm (Aslak & Alessandretti, 2020)
- where some time was spent

building a network

connected if a user had consecutive stops in two blocks within a day

community detection

using the network built from the stops
Louvain community detection is applied
- with different resolution values
- executed 10 times for each resolution

Louvain communities (resolution 2.5) — Louvain communities
(resolution 2.5)

Louvain community detection - resolution 2.5

infrastructural barriers: primary and secondary (dotted) roads

administrative boundaries: districts and neighborhood (dotted) — administrative boundaries:
districts and neighborhood (dotted)

barrier crossing ratio

$BCR_{\gamma}^{barrier} = \dfrac{1}{n} \frac{ \sum_{m} \text{{CB}} }{ \sum_{m} \text{{CB}} \times \text{{CC}}_{\gamma} }$

m is the number of mobility edges
$\gamma$ is the resolution
n is the number of iterations at resolution $\gamma$

by barrier types:

district
neighborhood
primary roads
secondary
railways
river

BCR by barrier types

classify users based on home location

trips within Budapest are considered

but the classification is not restricted to Budapest

decomposing barrier crossing ratio

back to amenities

$log BCR_{\gamma, n}^{c} = \alpha + \beta_1\log{D_c} + \beta_2\log{AC_j} + \epsilon$

$AC_j$ complexity of amenity portfolio of the visited $j$ location as proposed by (Juhász et al., 2023)

Nagoya metropolitan area

municipality boundaries wards (dotted) — municipality boundaries
wards (dotted)

open data (YJMob100K): (Yabe et al., 2024) | preprocessing (preprint): (Pintér, 2024)

BCR × Nagoya

takaway

there is a barrier effect
which affects people differently based on background

future work

socio-economic status
different time interval
- workdays - weekends
- time of the day

thanks for the attention!

Gergő Pintér, gergo.pinter @ uni-corvinus.hu, @pintergreg

this presentation is available online: pintergreg.github.io/ccs24

we are looking for contributors
with compatible data

already available on arXiv:
2312.11343

references

Aslak, U., & Alessandretti, L. (2020). Infostop: Scalable stop-location detection in multi-user mobility data. https://arxiv.org/abs/2003.14370

Juhász, S., Pintér, G., Kovács, Á. J., Borza, E., Mónus, G., Lőrincz, L., & Lengyel, B. (2023). Amenity complexity and urban locations of socio-economic mixing. EPJ Data Science, 12(1), 34.

Pintér, G. (2024). Revealing urban area from mobile positioning data. https://arxiv.org/abs/2407.18086

Yabe, T., Tsubouchi, K., Shimizu, T., Sekimoto, Y., Sezaki, K., Moro, E., & Pentland, A. (2024). YJMob100K: City-scale and longitudinal dataset of anonymized human mobility trajectories. Scientific Data, 11(1), 397.

stop distribution of agglomeration dwellers