Swine Flu Simulated by Social Networks

http://www.mccormick.northwestern.edu/news/articles/507

As a result of the new Swine Flu scare, a McCormick professor created a computer simulation that shows how swine flu could potentially spread over a long and short distance. His models replicate the ones we saw earlier in this course with the letter-simulation game.Professor Brockmann discuses how he simulated human mobility by tracking money (remember Wheresgeorge.com ?)

In the video, the Professor said he used WheresGeorge.com as a model for human movement. By using the tracking of where money goes, you can simulate how swine flu could potentially travel across the universe. Well what if it was the other way round. Maybe the reason why swine flu can travel so far so quickly is because of the liquidity of money. Money constantly changes hands, whether clean or dirty. It is possible that modeling swine flu by human movements isn't the most ideal case, but rather just monitoring money.

By tracking the cash network of the United States, the swine flu could be better modeled, as cash not only represents the thing that moves and changes hands the most, but can also be a carrier for money as people with swine flu could easily spend their infected money at a convenience store. That convenience store can deposit the infected money in a bank, and that bank could transfer money to other banks, and so on and so forth. The Bank system and the tracking of small bills ($1-$20) would provide a more ideal model for the movement of swine flu.

http://rocs.northwestern.edu/projects/swine_flu/

This website provides more data and graphs about the spread of the swine flu. Another interesting thing to note in this collection of data is how it refers to a "worst case scenario" spread of the disesase. This reminds me of other assumptions we make when working in social networks (such as everyone will have two new friends...). Modeling actual situations outside of ideal or specific conditions is much more difficult, but these estimates and assumptions allow us to approximate or guess potential outcomes. As this is the worst case situation, anything additional that is done to contain or treat the disease will lower the numbers, which makes the data still useful in the sense that it is not expected to be any worse than *insert estimate here*. Also, as more ACTUAL data is presented they are able to make additional claims or adjust the estimate as needed so the data can influence the predictions throughout its collection. All in all the use of social networks to chart the spread of a disease is an interesting approach, particularly by considering the exchange of money.

When we first we taught about the Random Wiring Model in class, I thought it was a bad model of a social network. But now that I am thinking about the Swine Flu and how it can spread, I think the Random Wire Model is can be a good model for certain networks. I always thought that we are so connected in our globalized world today that the "few random long-range links" was not realistic because every node will have a lot of random long-range links. Though this might be true for a social network like our friendships on Facebook were we all have a few random friends in some far away places, for a geographically laid out network like our city and towns on Earth, it applies very well. If you think about it, only a few cities are connected to some long range node by airports and sea ports. Also it is only part of the city that has all those long range connections like in Chicago it is Midway and O'hare has those connections. A city can be broken up into sub parts and all the sub parts of the city are connected via the wire model to each other base plainly if the right next to each other geographically. In order to get from one sub part to the another sub part, you have to go through all the sub parts of the city that is between. So to apply it to a disease, since most people stay most of their day within a few mile of their home, an disease would spread on the short edges to the nodes right next to the contaminated node. An example would be if a worker in who lives in O'hare gets swine flu, it would spread by going to Rolling Meadows to Arlington Heights to Des Plaines to Skokie and then to Evanston. So even though we are very interconnected world today, it would take some time for a disease to travel to the towns or cities not directly connected by airports or sea ports.