Rename all UI vars as in example:
cluteringUI -> UIclustering

As other output are also in this format (plotXYZ,) and it is easier to read.

Add appropriate errors for:

• at least two symptoms have to be selected to cluster?
• check the other errors and do the in a user friendly manner

E.g.: "Please select at least two symptoms to generate ...."

Se warning, error and message functions. Also try and tryCatch.

Draw confidence intervals instead of or in addition to columns in graph B (the barplot) of plotTests function.

Change the height of the figure dynamically, depending on the number of symptoms selected. This requires changes in the renderPlot function(how large is the image) as well as the plotOutput function(how much HTML space it gets).

Fix error that arises when only one symptom is selected.

EM data in TV file is "richer" than the Excel template data.

Add/allow for extra columns in the Excel file with data. The user can choose which one of the columns (all possible as candidates in first implementation) will be used for faceting. The option of using none for faceting is also valid.

Remove the message
WORKING WITH DEMO DATA!
appearing in the first tab before demo data are loaded

Create a clustering visualization of the data.

The code for tha graphs exist, but has to be modified for the particular application.

@llaarraa : please copy-paste the relevant code in this thread and reassign it back to me.

The upper part of the graph is cut off (intensity of the symtoms appears only for large values - problem that appears when selecting all the symptoms, too many, the legend is alligned with the lower margin of the graph; the upper part is cut off )

Several tables should be added:
Table "1"
Rows: all the symptoms selected
Columns:

• Proportion of positive in Group1 (0.00)
• Positive / All in Group1 (102/300)
• Proportion of positive in Group2 (0.00)
• Positive / All in Group2 (102/300)
• P-value (of differences in proportions test) (0.00)
• Confidence interval (of differences in proportions test) (0.00 ; 0.00)
  Threshold value should be used to establish which symptoms are positive. A table should be generated for each of the measurement occasions.

Table "2"
Rows: all the symptoms selected
Columns:

• Median of G1
• IQR of G1
• Median of G2
• IQR of G2
• P-value (returned by Mann-Whitney non parametric test of median equality, found as wilcox.test() in R)
  The symptoms threshold value used is always 0.

Table "3", "4"
Versions of above tables, but without the groups (and obviously without the P-values and confidence limits).

Create graphs of logistic regresion depending (odds ratio) for different groups in the data.

Code for graphs exists.

@llaarraa : Please copy relevant code into thread and reassign back to me.

Add checkbox (unchecked by default) that triggers faceting by the selected grouping variable on the graph.

Make sure the threshold value set by user influences the dataFiltered.yn() fnuction, which passes symptoms present to other functions.

Check if everything works correctly after change (if any is needed).

Change the horizontal axis so that values left and right of the zero value on the "Pyramid" graph are positive.

With the current installations the functions saved in separate files are not found
could not find function "plotPropWithSymptoms"
could not find function "tablePropWithSymptoms"
could not find function "tableMediansWithSymptoms"
could not find function "tableAllWithSymptoms"

After selecting demo data in the app

Error in file.exists(tfn) : invalid 'file' argument

Add checkbox or other input to enable sorting of units on the graph according to their similarity.

Use the seriation package to sort the units on the graph by their similarity.

Add to Excel options for user to:

• choose plotting of "death" (or potentialy some other date) YES/NO
• choose from which column should the date be selected

This could also be just one option, depending on design.

Finish help content that will be used by roxygen at each of the functions.

Message that demo data is being used should be implemented at the sidebar and removed from specific tabs.

Add option to use the grouping variable on the horizontal axis of the plot. Use it as a factor.

Try changing it to a factor in the dataset passed to function - see if it affects any existing functionality.

Rewrite code: Add dot (.) in front of internal functions, so they are ignored in help files (check how this works first).

Add checking if folder exists for the "Symptoms.xls", tab for setting folders containing R code etc.

Errors arise when Measurement=0 and Insomnia selected as symptom:
'p' must have the same length as 'x' and 'n'
missing values and NaN's not allowed if 'na.rm' is FALSE
missing values and NaN's not allowed if 'na.rm' is FALSE

Might be something with missing values in the dataset. Check and make the code more robust.

Remove comments, make the arguments standardised, check if there are any dependencies on column order, standardize variable names... or an appropriate subset of these.

Add into PlotSymptoms.xlsm

Error in data.yn[measurements == selectedProportion, ] :
incorrect number of dimensions

Probably dropped data frames - check code.

In the PARAMETERS sheet of PlotTests. Only offer R valid colors.

Add feature to update data when uploaded file changes, as allowed by the function:

reactiveFileReader {shiny}

Before symptoms are selected in the graph tab an error notice appears: Error: object 'value' not found - fix so that instead of error a warning saying "Select one or more variables to analyze in the left sidebar"

Error in library(lubridate) : there is no package called ‘lubridate’

install.packages("SVGAnnotation", repos="http://www.omegahat.org/R", type="source")
Warning in install.packages :
package ‘SVGAnnotation’ is not available (for R version 3.0.1)
Installing package into ‘C:/Users/lara/Documents/R/win-library/3.0’
(as ‘lib’ is unspecified)
trying URL 'http://www.omegahat.org/R/src/contrib/SVGAnnotation_0.93-1.tar.gz'
Content type 'application/x-gzip' length 4625544 bytes (4.4 Mb)
opened URL
downloaded 4.4 Mb

• installing source package 'SVGAnnotation' ...
  ** R
  ** data
  ** inst
  ** preparing package for lazy loading
  Warning: package 'Cairo' was built under R version 3.0.2
  Warning: package 'XML' was built under R version 3.0.2
  in method for 'genEdgeIds' with signature '"Ragraph"': no definition for class "Ragraph"
  in method for 'genEdgeIds' with signature '"graphNEL"': no definition for class "graphNEL"
  Creating a generic function from function 'getEdgeInfo' in package 'SVGAnnotation'
  in method for 'getEdgeInfo' with signature '"graphNEL"': no definition for class "graphNEL"
  ** help
  *** installing help indices
  ** building package indices
  ** installing vignettes
  ** testing if installed package can be loaded
  *** arch - i386
  Warning: package 'Cairo' was built under R version 3.0.2
  Warning: package 'XML' was built under R version 3.0.2
  *** arch - x64
  Warning: package 'Cairo' was built under R version 3.0.2
  Warning: package 'XML' was built under R version 3.0.2
• DONE (SVGAnnotation)

The downloaded source packages are in
‘C:\Users\lara\AppData\Local\Temp\Rtmpg1AEwq\downloaded_packages’

library(devtools)
Warning message:
package ‘devtools’ was built under R version 3.0.2
install_github(username="crtahlin", repo="medplot", ref="crt")
Installing github repo medplot/crt from crtahlin
Downloading medplot.zip from https://github.com/crtahlin/medplot/archive/crt.zip
Error: server error: (500) Internal Server Error

install.packages("SVGAnnotation", repos="http://www.omegahat.org/R", type="source")
Warning in install.packages :
package ‘SVGAnnotation’ is not available (for R version 3.0.1)
Installing package into ‘C:/Users/lara/Documents/R/win-library/3.0’
(as ‘lib’ is unspecified)
trying URL 'http://www.omegahat.org/R/src/contrib/SVGAnnotation_0.93-1.tar.gz'
Content type 'application/x-gzip' length 4625544 bytes (4.4 Mb)
opened URL
downloaded 4.4 Mb

• installing source package 'SVGAnnotation' ...
  ** R
  ** data
  ** inst
  ** preparing package for lazy loading
  Warning: package 'Cairo' was built under R version 3.0.2
  Warning: package 'XML' was built under R version 3.0.2
  in method for 'genEdgeIds' with signature '"Ragraph"': no definition for class "Ragraph"
  in method for 'genEdgeIds' with signature '"graphNEL"': no definition for class "graphNEL"
  Creating a generic function from function 'getEdgeInfo' in package 'SVGAnnotation'
  in method for 'getEdgeInfo' with signature '"graphNEL"': no definition for class "graphNEL"
  ** help
  *** installing help indices
  ** building package indices
  ** installing vignettes
  ** testing if installed package can be loaded
  *** arch - i386
  Warning: package 'Cairo' was built under R version 3.0.2
  Warning: package 'XML' was built under R version 3.0.2
  *** arch - x64
  Warning: package 'Cairo' was built under R version 3.0.2
  Warning: package 'XML' was built under R version 3.0.2
• DONE (SVGAnnotation)

The downloaded source packages are in
‘C:\Users\lara\AppData\Local\Temp\Rtmpg1AEwq\downloaded_packages’

library(devtools)
install_github(username="crtahlin", repo="medplot", ref="Crt")
Installing github repo medplot/Crt from crtahlin
Downloading medplot.zip from https://github.com/crtahlin/medplot/archive/Crt.zip
Installing package from C:\Users\lara\AppData\Local\Temp\Rtmpg1AEwq/medplot.zip
Installing medplot
"C:/PROGRA1/R/R-301.1/bin/x64/R" --vanilla CMD INSTALL
"C:\Users\lara\AppData\Local\Temp\Rtmpg1AEwq\devtools29b4401c1c92\medplot-Crt"
--library="C:/Users/lara/Documents/R/win-library/3.0" --install-tests

• installing source package 'medplot' ...
  ** R
  ** inst
  ** preparing package for lazy loading
  Warning: package 'Cairo' was built under R version 3.0.2
  Warning: package 'gplots' was built under R version 3.0.2
  Warning: package 'shiny' was built under R version 3.0.3
  Warning: package 'scales' was built under R version 3.0.3
  Warning: package 'ggplot2' was built under R version 3.0.3
  Warning: package 'seriation' was built under R version 3.0.3
  Warning: package 'XML' was built under R version 3.0.2
  ** help
  *** installing help indices
  ** building package indices
  ** installing vignettes
  ** testing if installed package can be loaded
  *** arch - i386
  Warning: package 'Cairo' was built under R version 3.0.2
  Warning: package 'gplots' was built under R version 3.0.2
  Warning: package 'shiny' was built under R version 3.0.3
  Warning: package 'scales' was built under R version 3.0.3
  Warning: package 'ggplot2' was built under R version 3.0.3
  Warning: package 'seriation' was built under R version 3.0.3
  Warning: package 'XML' was built under R version 3.0.2
  *** arch - x64
  Warning: package 'Cairo' was built under R version 3.0.2
  Warning: package 'gplots' was built under R version 3.0.2
  Warning: package 'shiny' was built under R version 3.0.3
  Warning: package 'scales' was built under R version 3.0.3
  Warning: package 'ggplot2' was built under R version 3.0.3
  Warning: package 'seriation' was built under R version 3.0.3
  Warning: package 'XML' was built under R version 3.0.2
• DONE (medplot)

  runApp(file.path(path.package("medplot"), "shinyapp_symptoms2" ))
  Error: could not find function "runApp"

  library(shiny)
  Warning message:
  package ‘shiny’ was built under R version 3.0.3

  runApp(file.path(path.package("medplot"), "shinyapp_symptoms2" ))
  Error in path.package("medplot") : none of the packages are loaded

  library(medplot)
  Loading required package: Cairo
  Loading required package: gplots
  KernSmooth 2.23 loaded
  Copyright M. P. Wand 1997-2009

Attaching package: ‘gplots’

The following object is masked from ‘package:stats’:

lowess

Loading required package: RColorBrewer
Loading required package: scales
Loading required package: reshape2
Loading required package: ggplot2
Need help? Try the ggplot2 mailing list: http://groups.google.com/group/ggplot2.
Loading required package: seriation
Loading required package: XML
Loading required package: plyr

Attaching package: ‘medplot’

The following object is masked from ‘package:gplots’:

plotCI

Warning messages:
1: package ‘Cairo’ was built under R version 3.0.2
2: package ‘gplots’ was built under R version 3.0.2
3: package ‘scales’ was built under R version 3.0.3
4: package ‘ggplot2’ was built under R version 3.0.3
5: package ‘seriation’ was built under R version 3.0.3
6: package ‘XML’ was built under R version 3.0.2

runApp(file.path(path.package("medplot"), "shinyapp_symptoms2" ))
Warning: package ‘gdata’ was built under R version 3.0.2
gdata: Unable to locate valid perl interpreter
gdata:
gdata: read.xls() will be unable to read Excel XLS and XLSX files unless the 'perl=' argument is used to specify
gdata: the location of a valid perl intrpreter.
gdata:
gdata: (To avoid display of this message in the future, please ensure perl is installed and available on the
gdata: executable search path.)
gdata: Unable to load perl libaries needed by read.xls()
gdata: to support 'XLX' (Excel 97-2004) files.

gdata: Unable to load perl libaries needed by read.xls()
gdata: to support 'XLSX' (Excel 2007+) files.

gdata: Run the function 'installXLSXsupport()'
gdata: to automatically download and install the perl
gdata: libaries needed to support Excel XLS and XLSX formats.

Attaching package: ‘gdata’

The following object is masked from ‘package:stats’:

nobs

The following object is masked from ‘package:utils’:

object.size

Error in library(lubridate) : there is no package called ‘lubridate’

Add a checkbox "Repeated measures exist". Checked by default. If unchecked, the "measurement" variable selected on GUI is ignored and its selection hidden and a custom "measurement" variable is created, where the values for all units are the same (we have a single measurement occasion).

The units with multiple measurement (those with multiple lines in the data file) are filtered out of the set, since we do not know what to do with them. A warning is generated on the existence of such units and placed in an appropriate tab ("Errors"?).

Checkbox should be positioned between the Date and Measurement inputs on the sidebar.

Add a static picture to be generated with plotTests (PDF and/or JPG), additionally to the SVG picture.
Maybe could be done as a parameter in Excel file, that a user can set..

We might define such a function:

medplotOnline=function(){
runApp(file.path(path.package("medplot"), "shinyapp_symptoms2" ))}

than the user just needs to call within R to open the browser - seen something (probably) similar in the diversity paper. ANy better ideas/ or already shiny-bilt-in variants?

medplotOnline()

It is probably defined somwhere in the DESCRIPTION file - check in R instructions.

Or some other easily accesible folder. Can redo in Latex if better. (Now in Word.)

Implement annotations on the clustering graphs. Add multiple selectInput control to select variables to be used. Show only variables, that are not selected as "symptoms".

For code see code sent by Lara, already implementing this. (see part "annotation=data.frame...", uiOutput("clusteringUIAnnotation")).

Only grouping variable should be selected by default.

Add checking if patient IDs are unique:
-to Excel?
-to R

If not unique, a warning should be passed from R to RExcel.

At the moment error is output. Fix the output so that dataFiltered() is shown in a proper format.

In PlotTests.xlsm.

The user can choose (via data entry on the tab) what kind of intervals will be used on this tab for plotting data.

Firstly, the user chooses how many intervals he wishes to have.
Secondly he chooses the limits of these intervals.
The data gets a special tag describing to which interval it belongs. Graphs are generated for each of the tags separately.

The graphs are already in "existence", but have to be modified to fit the particular function.

@llaarraa : Please copy-paste the relevant code into this issue and reassign it back to me.

RCS, logistf: remove PersonID from the possible choices of the variables

what is downloadable? Should we make some of the results explicity downloadable. Or just say in the paper that the user can copy the outputs? We will have to mention it in the paper

Now we have a default option for positivity of the variables, which is ok.

I would like to add an option in which the user says that he/she is not interested in using a thresholding value, and would therefore want to use ONLY the features related to numerical variables

like: add a checkbox in which he says something like: the variables should not be categorized . If he selects this option some of the tabs are not usable anymore - gets a printed message in all the tabs where variables are categorized saying The analyses included in this tab are available for categorical data only.

IN general: we have to think about how to deal with this issue. THe type of data that we are using as demo data are such that it is quite natural to use them in both ways/ categorical and numerical, but in general this would not be the case. How to deal with this?

Also, a user might also simply add data which are from the start categorical - and do not need further thresholding - Maybe add an additional checkbox saying:

My variables (as imported) are categorical
My variables (as imported) are numerical but I would like to categorize them using a positivity threshold of ....
My variables are numerical and should not be categorized

Add catching of errors so that:

• trivial errors do not print on the GUI disturbing the user
• significant errors print on the GUI, but are in a form that user can understand and act upon

Change: Choose symptoms (left sidebar) in Choose variables to analyze. Would change everywhere: symptoms to variables (or other appropriate denomination??)

Add (or change the existing graph) to plot confindence intervals instead of horizontal bars. See the second graph on the distributions tab for an example implementation.

Add all the fields needed in the DESCRIPTION file

At the moment zero values don't plot on the graph. Add their ploting with a special character (perhaps * or something that plots in the middle of the circle) at the date of the occurrence.