Eclipse Community Forums: Spatiotemporal Epidemiological Modeler (STEM)

Home » Eclipse Projects » Spatiotemporal Epidemiological Modeler (STEM) » Working with CSV files(Using CSV files to estimate case numbers)

Show: Today's Messages :: Show Polls :: Message Navigator

Working with CSV files [message #655245]

Fri, 18 February 2011 19:22

Tom Schafer

Messages: 5
Registered: February 2011

Junior Member

I am doubtlessly illustrating my ignorance here, but can someone please explain how the numbers derived from the CSV logger can be used to derive total case numbers for a given disease? At first glance, the tables seem straightforward, but when examined more closely, they make no sense. For example, when I examine the file for "total population - 0" that is at the national level for a run involving the United States, the initial number listed is 943211. Obviously, this is not the total population of the U.S. or anything like it. There are similar problems for numbers at the state and county level. Obviously, this leads me to doubt the incidence and fatality numbers. I am new to STEM, although not to epidemiology. Can someone tell me what I am doing wrong?

Report message to a moderator

Re: Working with CSV files [message #655557 is a reply to message #655245]

Mon, 21 February 2011 18:23

Stefan Edlund

Messages: 127
Registered: July 2009
Location: IBM

Senior Member

Hi Tom,

you're correct, the documentation on the log file was missing one essential piece of information, I added it to this page:

http://wiki.eclipse.org/Logging_Data_to_Files

"OBSERVE: The numbers in the log file are not aggregated, so if you are running a scenario down to admin level 2 (counties in US), the data in the log files for admin level 1 and 0 should not be used."

Also, I strongly recommend you run using the Runge Kutta integration before you start logging, since Finite Difference is only approximate.

Regards,
/ Stefan

Report message to a moderator

Re: Working with CSV files [message #657441 is a reply to message #655245]

Wed, 02 March 2011 19:12

James Kaufman

Messages: 240
Registered: July 2009

Senior Member

Tom,
Stefan and I have been investigating this.
The higher level nodes you were looking at reflect ONLY people in the airtransportation network. We don't log or aggregate the state data with the logger (the analysis perspective does this or you can use R or Matlab).
However, we agree the air transport data log is confusing. We will check in a fix.

Report message to a moderator

Issue is resolved Re: Working with CSV files [message #657451 is a reply to message #655245]

Wed, 02 March 2011 19:53

James Kaufman

Messages: 240
Registered: July 2009

Senior Member

Tom,
We checked in a fix to this.
The confusing data was in fact duplicate data as people in the air travel system were being double counted (really bad). The confusing high level nodes are no longer logged and there is no longer any double counting of people.
The fix is in the code base and will show up on the next regular build of STEM.

Report message to a moderator

Re: Issue is resolved Re: Working with CSV files [message #662443 is a reply to message #657451]

Wed, 30 March 2011 16:48

James Kaufman

Messages: 240
Registered: July 2009

Senior Member

Tom,
One last thing - addressing your original question. The category in the log that should correspond to "case number" is always "incidence" in STEM.

Report message to a moderator

Previous Topic:	Help running STEM Headless
Next Topic:	Welcome New STEM Committer Dirk Reuter

Goto Forum:

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

]

Current Time: Tue Apr 16 22:41:00 GMT 2024

.:: Contact :: Home ::.

Breadcrumbs

Sign up to our Newsletter