model size and data

[megboyar]

Hello,

I have a model which consists of several modules, e.g.

1. A demographic aging chain
2. A disease progression model
3. A housing model

These each have large datasets. I am currently saving a .vdf file for each of the models with 20-75 data variables in each, for example:

--> demographic_inputs.vdfx contains variables
-> births[age,<b>***</b>] 2020, 2021, ..........2070
-> mortality[age,<b>***</b>] 2020, 2021, ..........2070
-> etc.

1. Is it more parsimonious for the model which links these modules to use these .vdfs as inputs rather than the import of each separate variable into the main model?

2. In running scenarios, for clients using Vensim published models, I am making the .vdfs available as follows for them to choose from as follows:

--> births_basecase.vdfx
--> births_lowfertility.vdfx
--> births_high_fertility.vdfx

I think in published models the user can change the data source and by putting settings to adding new data sources to the end of the list, changing between these can override the main demographic_inputs.vdfx for specific variables. (?) But is there a better way to do this? Or to upload a set of datasets as a bespoke scenario chosen by the user? (Or, I guess I could pre-set entire scenarios for override such as demographics_highgrowth.vdfx which would include high fertility, high migration, etc.....)

3. Changes (.cin) files only work for constants, right?

4. Is there a way in published models to see a choice of arrayed variables? E.g. if I have deaths[age,<b>***</b>] can the user readily see deaths["0",female] deaths["0",male] or SUM(deaths[age!,<b>***</b>!)] or do I need to create a separate variable for any variation of SUM?

5. Can users export results from Vensim Model reader? Either individual graphs as they work or the entire model variables? I guess in the latter case it's best that I create and format export files with specific output variables?

6. Finally, if I publish this model to an internet site, and have a comment that clicks to open a .txt file, (e.g. to show variable metadata such that the user can copy it into a word doc), with this open with the user's home computer text editor? Any other ideas for how to work this?

Thank you,
Sarah

[Administrator]

1. Is it more parsimonious for the model which links these modules to use these .vdfs as inputs rather than the import of each separate variable into the main model?

I don't really understand what you mean by this.

2. In running scenarios, for clients using Vensim published models, I am making the .vdfs available as follows for them to choose from as follows:

--> births_basecase.vdfx
--> births_lowfertility.vdfx
--> births_high_fertility.vdfx

I think in published models the user can change the data source and by putting settings to adding new data sources to the end of the list, changing between these can override the main demographic_inputs.vdfx for specific variables. (?) But is there a better way to do this? Or to upload a set of datasets as a bespoke scenario chosen by the user? (Or, I guess I could pre-set entire scenarios for override such as demographics_highgrowth.vdfx which would include high fertility, high migration, etc.....)

I don't really understand what you mean by this.

3. Changes (.cin) files only work for constants, right?

Correct.

4. Is there a way in published models to see a choice of arrayed variables? E.g. if I have deaths[age,<b>***</b>] can the user readily see deaths["0",female] deaths["0",male] or SUM(deaths[age!,<b>***</b>!)] or do I need to create a separate variable for any variation of SUM?

You'll need to create separate variables.

5. Can users export results from Vensim Model reader? Either individual graphs as they work or the entire model variables? I guess in the latter case it's best that I create and format export files with specific output variables?

Yes, you can export results from the model reader.

6. Finally, if I publish this model to an internet site, and have a comment that clicks to open a .txt file, (e.g. to show variable metadata such that the user can copy it into a word doc), with this open with the user's home computer text editor? Any other ideas for how to work this?

It will open using the user home computer editor.

[tomfid]

It's more parsimonious to load different data files at runtime, but you can also do something like the following:

scenario : low, mid, hi

data pop[scenario]

data pop selected = IF THEN ELSE(active scenario = low, data pop[low] ... etc.

Whether parsimonious matters is just a function of runtime.

[tomfid]

Re 4: there are some new I/O controls for subscript selection on screen, or people can use the Subscript Control. However, there isn't a way to do aggregation (SUM etc.) on the fly - you have to create the variables in advance.