Eclipse Community Forums: Spatiotemporal Epidemiological Modeler (STEM)

Help

Home

Home » Eclipse Projects » Spatiotemporal Epidemiological Modeler (STEM) » Earth science data file names changed

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

Earth science data file names changed [message #634239]

Wed, 20 October 2010 18:10

Stefan Edlund
Messages: 127
Registered: July 2009
Location: IBM

Senior Member

The file names used to store earth science data, for instance USA_2_earthscience_rainfall.graph has been renamed to include the valid year. So now the name is USA_2_earthsceience_2009_rainfall.graph

As a consequence, if you have any models containing earth science graphs, you need to either recreate them or manually edit the model file to add the missing year string.

This will take effect in the next weekly integration build after 10/20.

Regards,
/ Stefan

Report message to a moderator

Re: Earth science data file names changed [message #644510 is a reply to message #634239]

Sun, 12 December 2010 10:34

Matthias Filter
Messages: 73
Registered: July 2009

Member

Hi,
I just had a look on the GlobalEarthSience Demo Example and there are several issues that I noticed:
- in the Europe models the dublin core Identifier of the node labels of vegetation / temperature and elevation alway refers to rainfall, in the NorthAmerica Model all the identifiers are set to "elevation"
- in case of the rainfall nodes there are several times list values of "99999.0" given which probably means some kind of NA or not available. I think this should be explained somewhere or these values should be removed.
- there is no way in the property view to edit or even look at the whole list of values for a node. As there are so many values given for each node the screen is not wide enough to even look at them.
- most important: at no place there is an information given on how to interprete the values given in the nodes Data attribute. E.g. it is not clear whether the numbers refer to days, weeks or month of a year

Concerning the later use of these models I would be interested in knowing whether there is a way around the fact that it seems to be necessary to implement always a specific disease models to make use of these data. E.g. I could imagine that one would want to use the specific rainfall or vegetation value at a given simulation time (e.g. the corresponding value for each week of the year 2009) instead of an averaged value for a given location. Do you have also an idea on that?
Matthias

Report message to a moderator

Re: Earth science data file names changed [message #645891 is a reply to message #634239]

Mon, 20 December 2010 19:46

Stefan Edlund
Messages: 127
Registered: July 2009
Location: IBM

Senior Member

Hi Matthias,

to answer your questions:

1. I think the rainfall data is missing for countries above a certain latitude, so a large parts of europe is missing. There's another defect (314326) that might also explain why dublin core isn't refreshed in the properties view.

2. Yes, 99999.0 indicates missing data. The documentation of what the values represent are in the original .properties files. Unfortunately, the text is just comments inside those files and is not carried over to the actual graphs we build. It would be possible to regenerate the graphs putting the documentation inside the dublin core description field (for the graph) instead so they become available to the user. Would that work?

3. Agree, this is a problem. Would this be solved once we have a full fledged graphical editor?

4. Yes, see 2.

Report message to a moderator

Re: Earth science data file names changed [message #645893 is a reply to message #634239]

Mon, 20 December 2010 19:54

Stefan Edlund
Messages: 127
Registered: July 2009
Location: IBM

Senior Member

Also, regarding the use of the earth science data. We have only one example of where we use it right now, and it's in a a population model that estimates mosquito density.

The environmental data are monthly averages for each region (as well as max/min for each data point inside the region) and only for the year 2009. This is a limitation, but at least it's a start. For our mosquito density model we do some interpolation to get a value for a particular day in the month. The date we use comes from the sequencer, but we ignore the year (it's always assumed 2009).

Report message to a moderator

Re: Earth science data file names changed [message #646767 is a reply to message #634239]

Thu, 30 December 2010 15:35

Stefan Edlund
Messages: 127
Registered: July 2009
Location: IBM

Senior Member

The dublin core Description property (for the earth science graph) now has a documentation for what the field values represent.

Also, the earth science label value contains a property 'Units'. For rainfall, it's mm, for elevation it's m etc.

Regards,
/ Stefan

Report message to a moderator

Previous Topic:	JUnit Tests
Next Topic:	User opinions re: Default Map Background Color

Goto Forum:

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

]

Current Time: Sat May 25 17:32:52 EDT 2013