Network epidemic model and App Movement

2017-04-21

Connected classes

User / cumulative count	Share link	Time created	Parent user	Parent share link
1	A	2017-04-21 00:00:00
2	B	2017-04-21 00:39:26	1	A
3	C	2017-04-21 05:59:05	1	A
4	D	2017-04-21 09:44:41	2	B
5	E	2017-04-21 13:32:37	4	D
6	F	2017-04-21 14:03:21	4	D
7	G	2017-04-21 19:36:18	6	F
8	H	2017-04-21 20:18:07	3	C
9	I	2017-04-21 21:45:21	4	D
10	J	2017-04-22 01:02:49	2	B
11	K	2017-04-22 06:55:49	2	B
12	L	2017-04-22 08:20:30	11	K

A few parameters to explain a relationship or phenomenon

One parameter needed: the epidemic rate

the simplest version of epidemic models is called the susceptible-infectious model, where you start with one person being infectious, and infects anybody that comes into contact with that person
those infected become infectious and will infect others who are still susceptible
this is plotting the cumulative count of the infected and susceptible over time
the epidemic slows down towards the end because there are not many people left to be infected
for app movement, we can define anybody who hasn't seen the campaign page as susceptible. Once they click the link shared by others on the social network, they become infected.
in this way, this graph is telling the same story as this table (go back 2 slides)
all you need to create this epidemic is the constant rate of contacting other people
if any of you knows about epidemiology, this is actually related to the basic reproduction number

Everyone is connected to everyone else

Everyone has the same average number of connections

Some are much more connected than others - the hubs

Start with one connection

Newcomers sequentially connect to those already in the network

Current connectedness affects how likely one gets connected to a newcomer

The rich gets richer

The epidemic rate kicks in

Most connections are unknown