Social trajectories

Binned

The term social trajectory was coined by Sekara, Stopczynski, and Lehmann and refers to the groups an individual node was part of over time.

In their study they used a binned social trajectory, showing the groups each node was part of each day.

To reproduce this, use the function tacoma.api.binned_social_trajectory()

It returns a set of group indices the single node was part of in each time interval.

Here’s an example binned for each hour.

from tacoma.analysis import plot_social_trajectory
import matplotlib.pyplot as pl

ht09 = tc.load_sociopatterns_hypertext_2009()
binned_traj = tc.binned_social_trajectory(ht09, node=1,
                                          N_time_steps=int(ht09.tmax/3600.))

fig, ax = pl.subplots(1, 1, figsize=(4,3))
for it, groups in enumerate(binned_traj):
    for g in groups:
        ax.plot([it,it+1], [g,g], 'k')
ax.set_xlabel('hour')
ax.set_ylabel('group node 1 was part of')
pl.show()

Binned social trajectory for node 1, showing the groups of size \(g>1\) it was part of each hour (Sociopatterns Hypertext 09 dataset

Alternatively, use the function tacoma.analysis.plot_binned_social_trajectory().

Continuous time

A more sensible function respecting continuous time is the pure social trajectory returned by tacoma.api.social_trajectory().

It can be easily computed and plotted using tacoma.analysis.plot_social_trajectory().

from tacoma.analysis import plot_social_trajectory
import matplotlib.pyplot as pl

ht09 = tc.load_sociopatterns_hypertext_2009()
soc_traj = tc.social_trajectory(ht09, node=1)

fig, ax = pl.subplots(1, 1, figsize=(4,3))
plot_social_trajectory(soc_traj, ax, time_unit='s')
pl.show()

Social trajectory for node 1, showing the groups of size \(g>1\) it was part of (Sociopatterns Hypertext 09 dataset