statnet/statnetweb

A graphical user interface for network modeling with the 'statnet' software

statnet: Software Tools for the Statistical Analysis of Network Data

An integrated set of tools for the representation, visualization, analysis, and simulation of network data. For an introduction, type help(package='statnet').

Public and Private repositories

To facilitate open development of the package while giving the core developers an opportunity to publish on their developments before opening them up for general use, this project comprises two repositories:

• A public repository statnet/statnet
• A private repository statnet/statnet-private

The intention is that all developments in statnet/statnet-private will eventually make their way into statnet/statnet and onto CRAN.

Developers and Contributing Users to the Statnet Project should read https://statnet.github.io/private/ for information about the relationship between the public and the private repository and the workflows involved.

statnetweb's Issues

Problem with fitting nodecov example on flomarriage from NME Day 2

This is with the CRAN version running locally, here is the error message under the hood.

Warning in strsplit(msg, "\n") :
  input string 1 is invalid in this locale
Error in parse(text = x, keep.source = FALSE) : <text>:1:21: unexpected input
1: nw() ~ edges+nodecov(�

Web version of statnetWeb MCMC diagnostics not loading lattice plots correctly

CRAN error

CRAN folks found the following error and have asked us to fix it:

Found the following \keyword or \concept entries
which likely give several index terms:
  File 'run_sW.Rd':
    \keyword{GUI,}

The offending line is here

Any idea what they're asking for?

Define degree and geodesic distribution in Network descriptives section

It would be helpful to have a little description for the degree distribution and geodesic distribution tabs explaining these metrics, especially for the latter where the Inf component of the distribution is not intuitive. Perhaps these could be a little ? click help window.

Geodesic distance counts

Geodesic distance distribution count exceeds total possible dyad counts (e.g., for faux.mesa.high: choose(205, 2) = 20910)

link to 'commonly used ergm terms' on Fit Model > Term Documentation is broken

http://statnet.csde.washington.edu/EpiModel/nme/d2-ergmterms.html just gives 404

I think sam moved things to github io account?

clear model fit iterations and summary after new network is uploaded

currently the summary statistics get updated, but before clicking 'Fit Model' again, the model fit from previous network is still visible

Type in terms to use for cug histograms

selectizeInput is a good format to use for typing/searching from a list, but some terms have additional arguments and some don't. Also, terms differ in the form of their outputs, and the histograms are only set up for using terms with single values outputs.

interactive network plot

Prefit summary

1. Retitle as "Summary Statistics"
2. If possible, use more output columns (currently uses only 2)

Factor ordering for legends and summary model output

Current ordering is a bit odd. For example, look at faux.mesa. Here, alphabetic (with "other" at end) would be best, but may be difficult to achieve. Note that the same ordering would be useful for nodefactor and nodematch in the model output. If R is using the order encountered in the data, sorting might help, but: 1. i'm not sure how to sort a network object, and 2. achieving alphabetical order for multiple factors may not be possible; and 3. that wouldn't guarantee "other" at the end.

Worth looking at the "relevel" command -- maybe that will work here.

unicode characters in documentation box

For systems that are not UTF-8 based (Windows), be explicit about encoding so that unicode characters display properly

ecoli1 and 2 networks generate errors

Error in solve.default(A, tol = tol) :
system is computationally singular: reciprocal condition number = 1.7338e-21

Can others please try pulling to make sure you're working with the same version, and reproduce this. The error shows up on the console when you load the network from the Data tab

change formatting for simulation summary tab

Parameters should be listed by row, not across columns

Ditto the network stats from multiple sims, the rows should be the stats, the cols should be the sims

shinyapps: cannot load ecoli data sets

"Calc in progress" popup

1. it doesn't seem to pop up in the shiny window on my windows pc, but it does show up when i view in browser
2. might be worth making it pop up in the active viewing area -- right now it's only visible if you scroll to the top of the screen

use network name uploaded by user as object name

Functions within the shiny code that take in a network object use the nw() reactive object, which is convenient, except for a few scenarios when a different behavior would be more helpful for the user.

• When fitting a model, edgecov and dyadcov terms require nw() argument instead of actual network name
• The printed model summary shows nw() ~ edges

This is probably a long term issue.

minor issue with "What is an R-object?" help text

I'm pretty sure the help text for "What is an R-object?" is wrong about setting the path to the new location the 'envir' argument (which controls the internal search path for the object). User should either include the full path when giving the file name, or change R's working directory. Or set file=choose.file(new=TRUE) to use a normal save dialog.

add MCMC control parameters to fitting interface

Ideally, in a set of fitting option tabs on the upper right

error downloading degree dist & geodesic dist plots

due to replacing the input$densplot variable with the state$plotperc_dd reactive value (and corresponding values for the geodesic plot)

Error messages in statnetWeb: fixing before SunBelt

@martinamorris @chad-klumb

This is just an issue to get us started on documenting errors and fixes.

I've been able to get the app running and have successfully (perhaps unfortunately) the same errors Martina showed us.

Here is the list:

• Chad to look into the "summary.formula" to fix this error:

This can be recreated: Data -> upload network -> filetype (built in network is faux.mesa.high) ; Network Descriptives -> More -> Run the default term on 100 simulations.

I'll be away from my laptop till 7PM EST (4PM PST) but will bug test this out as much as possible when I return

disable widgets when not applicable

• choosing directed/undirected for after symmetrizing
• degree type for degree distribution

clean up appearance

• solidify color scheme of action buttons
  • gradient on hover?
• helper buttons should line up with what they explain
• tool tips should not obscure what they explain (see statnet Wiki link in resources panel)
• color of "Calculation in Progress" icon
• delay before tool tips appear?

(some work already done in commits 87cb384, 75411bd, a732502)

Package structure revision needed to deploy

Shiny app package structure requirements have changed since this was first written. When I try to deploy to the shinyApps.io server I get the following error:

deployApp()

Error in lint(appDir, appFiles, appPrimaryDoc) :
Cancelling deployment: invalid project layout.
The project should have one of the following layouts:

1. 'shiny.R' and 'ui.R' in the application base directory,
2. 'shiny.R' and 'www/index.html' in the application base directory,
3. 'app.R' or a single-file Shiny .R file,
4. An R Markdown (.Rmd) document,
5. A static HTML (.html) or PDF (.pdf) document.
6. 'plumber.R' API description .R file
7. 'entrypoint.R' plumber startup script
8. A tensorflow saved model

clear GOF plot window when new model is run

otherwise, the plots from the old model are there, and may be mistaken for new plots

deployApp error log

for some reason, it's not finding the ergm datasets.

> rsconnect::showLogs()
2019-06-15T19:17:25.218471+00:00 shinyapps[42335]: 
2019-06-15T19:17:25.245307+00:00 shinyapps[42335]: Listening on http://127.0.0.1:39309
2019-06-15T19:17:25.245304+00:00 shinyapps[42335]: 
2019-06-15T19:17:32.095261+00:00 shinyapps[42335]: Warning in if (!is.na(attribValue)) { :
2019-06-15T19:17:32.095264+00:00 shinyapps[42335]:   the condition has length > 1 and only the first element will be used
2019-06-15T19:17:32.095419+00:00 shinyapps[42335]: Warning in charToRaw(enc2utf8(text)) :
2019-06-15T19:17:32.095420+00:00 shinyapps[42335]:   argument should be a character vector of length 1
2019-06-15T19:17:32.095421+00:00 shinyapps[42335]: all but the first element will be ignored
2019-06-15T19:17:34.088237+00:00 shinyapps[42335]: Warning in data(faux.mesa.high) : data set ‘faux.mesa.high’ not found
2019-06-15T19:17:34.093196+00:00 shinyapps[42335]: Warning in data(florentine) : data set ‘florentine’ not found
2019-06-15T19:17:34.097437+00:00 shinyapps[42335]: Warning in data(sampson) : data set ‘sampson’ not found
2019-06-15T19:17:34.101819+00:00 shinyapps[42335]: Warning in data(samplk) : data set ‘samplk’ not found
2019-06-15T19:17:34.106127+00:00 shinyapps[42335]: Warning in data(ecoli) : data set ‘ecoli’ not found
2019-06-15T19:17:34.110526+00:00 shinyapps[42335]: Warning in data(molecule) : data set ‘molecule’ not found
2019-06-15T19:17:34.114905+00:00 shinyapps[42335]: Warning in data(kapferer) : data set ‘kapferer’ not found
2019-06-15T19:17:34.128370+00:00 shinyapps[42335]: Warning: Error in search.ergmTerms: could not find function "search.ergmTerms"
2019-06-15T19:17:34.145717+00:00 shinyapps[42335]:   62: splitargs [/srv/connect/apps/statnetWeb/global.R#40]
2019-06-15T19:17:34.145897+00:00 shinyapps[42335]: Error in search.ergmTerms(keyword = searchterm, net = nw) : 
2019-06-15T19:17:34.145899+00:00 shinyapps[42335]:   could not find function "search.ergmTerms"
2019-06-15T20:57:14.230891+00:00 shinyapps[42335]: Server version: 1.7.2-7
2019-06-15T20:57:14.230937+00:00 shinyapps[42335]: LANG: en_US.UTF-8
2019-06-15T20:57:14.449107+00:00 shinyapps[42335]: Starting R with process ID: '19'
2019-06-15T20:57:14.230941+00:00 shinyapps[42335]: R version: 3.6.0
2019-06-15T20:57:14.231876+00:00 shinyapps[42335]: Using pandoc at /opt/connect/ext/pandoc2
2019-06-15T20:57:14.443564+00:00 shinyapps[42335]: Using jsonlite for JSON processing
2019-06-15T20:57:14.449103+00:00 shinyapps[42335]: 
2019-06-15T20:57:14.230944+00:00 shinyapps[42335]: shiny version: 1.3.2
2019-06-15T20:57:14.230945+00:00 shinyapps[42335]: httpuv version: 1.5.1
2019-06-15T20:57:14.231010+00:00 shinyapps[42335]: knitr version: NA
2019-06-15T20:57:14.231015+00:00 shinyapps[42335]: htmltools version: 0.3.6
2019-06-15T20:57:14.231013+00:00 shinyapps[42335]: jsonlite version: 1.6
2019-06-15T20:57:14.231014+00:00 shinyapps[42335]: RJSONIO version: NA
2019-06-15T20:57:14.230946+00:00 shinyapps[42335]: rmarkdown version: NA
2019-06-15T20:57:14.482240+00:00 shinyapps[42335]: 
2019-06-15T20:57:14.482243+00:00 shinyapps[42335]: Listening on http://127.0.0.1:35505
2019-06-15T20:57:20.138268+00:00 shinyapps[42335]:   the condition has length > 1 and only the first element will be used
2019-06-15T20:57:20.138390+00:00 shinyapps[42335]:   argument should be a character vector of length 1
2019-06-15T20:57:20.138392+00:00 shinyapps[42335]: all but the first element will be ignored
2019-06-15T20:57:20.138388+00:00 shinyapps[42335]: Warning in charToRaw(enc2utf8(text)) :
2019-06-15T20:57:20.138245+00:00 shinyapps[42335]: Warning in if (!is.na(attribValue)) { :
2019-06-15T20:57:22.868502+00:00 shinyapps[42335]: Warning in data(faux.mesa.high) : data set ‘faux.mesa.high’ not found
2019-06-15T20:57:22.872337+00:00 shinyapps[42335]: Warning in data(florentine) : data set ‘florentine’ not found
2019-06-15T20:57:22.876235+00:00 shinyapps[42335]: Warning in data(sampson) : data set ‘sampson’ not found
2019-06-15T20:57:22.884102+00:00 shinyapps[42335]: Warning in data(ecoli) : data set ‘ecoli’ not found
2019-06-15T20:57:22.888045+00:00 shinyapps[42335]: Warning in data(molecule) : data set ‘molecule’ not found
2019-06-15T20:57:22.891947+00:00 shinyapps[42335]: Warning in data(kapferer) : data set ‘kapferer’ not found
2019-06-15T20:57:22.905992+00:00 shinyapps[42335]: Warning: Error in search.ergmTerms: could not find function "search.ergmTerms"
2019-06-15T20:57:22.922676+00:00 shinyapps[42335]:   62: splitargs [/srv/connect/apps/statnetWeb/global.R#40]
2019-06-15T20:57:22.880179+00:00 shinyapps[42335]: Warning in data(samplk) : data set ‘samplk’ not found
2019-06-15T20:57:22.922871+00:00 shinyapps[42335]:   could not find function "search.ergmTerms"
2019-06-15T20:57:22.922869+00:00 shinyapps[42335]: Error in search.ergmTerms(keyword = searchterm, net = nw) : 

Online version of statnetWeb can't replicate NME Day 2 add health four model comparison

Specifically, in the third model with just attribute mixing (without gwesp), this error message when doing the goodness of fit statistics

Error: coef has 20 elements, while the model requires 19 parameters.

Summary and Fit tabs have intermingled output

When I refit a new model with the summary tab active, the fit output goes to the summary tab, as well as the summary output (using view in browser).

Default plot margins in Network Descriptives section

To get the network plot to better fill the available plot window in the Network Plot tab under the Network Descriptives section, try setting the margins for the plot window to 0 on all four sides:

par(mar = c(0, 0, 0, 0))

That might look better.

error downloading cugtest plots

The file name appears in the downloads folder, but has no contents. The following error prints to the console:

Listening on http://127.0.0.1:6930
Error opening file: 2
Error reading: 6

This discussion addresses the same issue, but their solution isn't working.

library refs in server.R

Just a minor note on loading libraries in your inst folder files.

Now that you have all these packages as depends, this:
https://github.com/statnet/statnetWeb/blob/master/inst/shiny/server.R#L2-L8

can just be replaced with:

library(statnetWeb)

One thing that you may have found out from trying to deploy to shinyapps.io is that it will not accept a package that is built locally, but it will accept one built from github. Therefore, after you commit to github, you need to:

1. Install statnetWeb with `devtools::install_github("statnet/statnetWeb")
2. Then deploy

missing library dependency with default instructions

When I ran the default command after installing the shiny library I got the following error

shiny::runGitHub("statnetWeb", "statnet")
Downloading https://github.com/statnet/statnetWeb/archive/master.tar.gz
Loading required package: shiny

Listening on http://127.0.0.1:7254
Error in library(shinyBS) : there is no package called ‘shinyBS’

After I manually installed shinyBS it worked fine. Not sure if the instructions need to say install shinyBS, or if the app is missing a dependency declaration or something?

add conditional uniform graph tests

Hopefully utilizing the cug.test function in sna

add sortable table of attribute values

Transition online app to statnet.shinyapps.io

@ebey We just purchased the next tier of shinyapps service for a year, so when you're ready let's move your statnetWeb app over to our statnet shinyapps account. That will support many more user hours per month.

Loading a new network dataset after having clicked on a vertex gives "Undifiend columns" error

If I load one dataset, click on a vertex in the network plot, and go load another dataset, when I return to the plot it gives an error. The error goes away after clicking again or refreshing the plot

Plotting legend

fix colors

clearer first step when arriving in the app?

I think the initial landing About v0.1.0 page should have somewhere a big "Start Here" button that jumps to the first page in the tab sequence. Otherwise you have to read the text pretty carefully to figure out the 1st step. (or do what I did, and click on the statnet citation because it looks like a button :-) This is especially true if you happen to be viewing on a narrower screen, as the responsive layout will (nicely) collapse the tabs into a popup menu.

needs some minimal tests

writing tests for a GUI application can be quite hard, but it would be great to have some minimal tests that the application can launch so we know if we break it with upstream changes

MCMC diagnostics section for dyadic independent models

When a model containing only dyadic independent terms is fit (for example, ~edges + nodematch("Grade") for the add health datasets), no MCMC is run as expected. But it may be helpful for the novice user to have some explanation of this in the MCMC diagnostics tab with such models.

Another option is to always force MCMC even with dyadic independent models with the control.ergm setting force.main set to TRUE, which would always run MCMC. But I'm not sure if this is a good statistical practice.

attribute histogram options

would be nice to have option to plot Y axis as percent or n

Error : Invalid DESCRIPTION file

Installing statnetWeb gives the following error.

devtools::install_github("statnet/statnetWeb")

* installing *source* package 'statnetWeb' ...
Warning in .canonicalize_person_role(role) :
  Invalid role specification: 'cre, aut'.
Error : Invalid DESCRIPTION file

Authors@R field gives no person with maintainer role, valid email
address and non-empty name.

See section 'The DESCRIPTION file' in the 'Writing R Extensions'
manual.

ERROR: installing package DESCRIPTION failed for package 'statnetWeb'
* removing 'F:/ROpen/R-3.5.1/library/statnetWeb'
In R CMD INSTALL
Installation failed: Command failed (1)

v0.3.0 not deployed to shinyapps

shinyapps does not currently support the Rglpk package, which is used in the newest version of ergm.

nodefactor terms for multiple factors

Objective: allow user to pick nodefactor terms for multiple factors independently

Current behavior: picking multiple factors results in nodefactor(c(x1, x2)) instead of the desired nodefactor(x1) + nodefactor(x2)

clear GOF plot window when new model is run

otherwise, the plots from the old model are there when you flip to that tab. the MCMC do eventually redraw automatically, but the GOF don't, and may be mistaken for new plots

add brief description of built-in datasets within app

make all plots reactive objects

The reactive object will get passed to both renderPlot to show to the user and downloadHandler if they choose to download it, saving run time.

Also, consider changing image formats from PDF to PNG.

histogram/table of attribute totals

gather errors from centrality measures

When calculating graph or node-level centrality measures that return an error (e.g. information centrality on the ecoli1 network), save the error in a reactiveValues list and output the list in box at the bottom right of the screen. Link the error message to the place in the app where the error occurred with an error number. Ideally, we will eventually show the user all errors in this way.

Currently, the framework is set up just for the node-level information centrality measure, but the conditionalPanel for the error message box is not working as expected. Fix this first, then work on adding more errors to the list.

how to generate help file html

http://stat.ethz.ch/R-manual/R-devel/library/base/html/RdUtils.html