Storm Tracker | Mobility Impact | Storm Stats | City Boundaries | Data Processing | Impact Charts | Refinements

Mobility Impact Charts

Project Objective: Assess the degree to which the maximum sustained wind speed on the day of storm impact correlates with mobility pattern changes (daily displacement and radius of gyration) using geotagged Twitter posts.

Perturbation Strengths

Impact of Wind on Tweet Displacement Distances

Radius of Gyration

Radius of Gyration - Cities

Radius of Gyration in kilometers for individuals moving between their Twitter posts. Zero indicates the day of storm impact.

2+ locations
3+ locations
5+ locations

* Normalized images have not yet been added for 3-day segments, nor 5+ locations.


A single pixel line on the 0 y-value indicates no movement data for the day for the selected number of stop minimum (2, 3 or 5 stops).
For some days there is no data due to data collection outages.


Radius of Gyration - Tacloban

Scale of y-axis changed from 40 to 4 km to show more detail for Tacloban, which has an 11 km radius.


Tacloban chart (2+ stops): The current Seaborn graph process requires setting the radius to 0
which produces a line regardless of whether there is any movement data for the day.

More Radius of Gryation Charts


Calculating Distance on the Earth's Spherical Surface

The angles of a triangle are more than 180° when applied to a sphere.



Radius of Gyration on a 2D surface:


Radius of Gyration on 3D surface:

Haversine Formula (Robusto 1957)



In PostgreSQL, the distance between two points on earth in kilometers can be calculated using ST_DistanceSphere:

displacement = ST_DistanceSphere(prior_geom, the_geom)/1000
ST_DistanceSphere is equivalent to the following implementation of the Haversine Formula (with less than 1 meter variance):

displacement_haversine = 2 * 6371 * asin(sqrt((sin(radians((latitude - prior_latitude) / 2))) ^ 2 
+ cos(radians(prior_latitude)) * cos(radians(latitude)) * (sin(radians((longitude - prior_longitude) / 2))) ^ 2))
Extra Charts