The current state of the app is considered experimental, and a lot of things are still changing. It is, however, in a usable state, with basic functions available to the user.
-
Below are some (the actual list is quite long and growing) of the planned features and improvements:
-
-
Additional study designs in regression models (expansion of the regression analysis functionality have been put on hold for now to focus on the more basic use-cases):
+
The current state of the app is considered experimental, however, from version 25.4.2, the FreesearchR app is considered functional and can be used for data evaluation and analyses.
+
Below are some (the actual list is quite long and growing) of the planned new features and improvements:
+
+
Implementation in real world clinical studies and projects
+
This really is the main goal of the whole project.
+
+
+
New features:
+
+
+
+
+
+
Expanded options for regression models:
+
Expansion of the regression analysis functionality have been put on hold for now to focus on the more basic use-cases):
+
More study designs
-
-
-
-
More detailed variable browser
-
-
-
-
-
More output controls
-
-
-
-
-
-
-
-
-
Graphs and plots
-
-
-
-
-
Documentation:
-
-
-
New features:
-
-
+
Other regression models
+
+
+
+
Data handling
+
+
+
Improved documentation:
+
+
+
-
+
+
Data
+
+
This is the panel to get a good overview of your data, check data is
+classed and formatted correctly, perform simple modifications and filter
+data.
+
+
Summary
+
+
Here, the data variables can be inspected with a simple visualisation
+and a few key measures. Also, data filtering is available at two
+levels:
+
+
Data type filtering allows to filter by variable data
+type
+
Observations level filtering allow to filter data by
+variable
+
+
+
+
Modify
+
+
+
+
Evaluate
This panel allows for basic data evaluation.
@@ -219,29 +249,6 @@ visualise data distribution
set relationships
-
ordinal
-
Stacked horizontal bars
-
A classical way of visualising the distribution of an
-ordinal scale like the modified Ranking Scale and known as Grotta
-bars
-
-
-
ordinal
-
Violin plot
-
A modern alternative to the classic boxplot to
-visualise data distribution
-
-
-
ordinal
-
Sankey plot
-
A way of visualising change between groups
-
-
-
ordinal
-
Box plot
-
A classic way to plot data distribution by groups
-
-
categorical
Stacked horizontal bars
A classical way of visualising the distribution of an
@@ -277,6 +284,41 @@ fine tune all the small details.
This section is only intended for very simple explorative analyses
and as a proof-of-concept for now. If you are doing complex regression
analyses you should probably just write the code yourself.
+
Below are the available regression types listed.
+
+
+
+
+
+
+
+
+
Data type
+
Regression model
+
Function
+
Study design
+
+
+
+
continuous
+
Linear regression model
+
stats::lm
+
cross-sectional
+
+
+
dichotomous
+
Logistic regression model
+
stats::glm
+
cross-sectional
+
+
+
categorical
+
Ordinal logistic regression model
+
MASS::polr
+
cross-sectional
+
+
+
Table
@@ -305,7 +347,7 @@ Choose which model(s) to include.
model results. Choose between MS Word or LibreOffice format.
+library(FreesearchR)
+#>
+#> Attaching package: 'FreesearchR'
+#> The following object is masked from 'package:base':
+#>
+#> sort_by
+
+
A clinical data class
+
+
Traditionally in R, data is identified by classes, like
+numeric, integer, double, logical, factor etc. These classes can be a
+little confusing from a clinical or operational standpoint. In the
+FreesearchR app, these classes has been
+simplified and modified to the following data types, that are assigned
+on a prioritised order like the following:
+
+
+
+
+
+
+
+
Data type
+
Description
+
Data classes included
+
+
+
+
empty
+
Variable of all NAs
+
Any class
+
+
+
monotone
+
Variable with only one unique value
+
Any class
+
+
+
dichotomous
+
Variable with only two unique values
+
Any class
+
+
+
categorical
+
Factor variable
+
factor (ordered or unordered)
+
+
+
text
+
Character variable
+
character
+
+
+
datetime
+
Variable of time, date or datetime values
+
hms, Date, POSIXct and POSIXt
+
+
+
continuous
+
Numeric variable
+
numeric, integer or double
+
+
+
unknown
+
Anything not falling within the previous
+
Any other class
+
+
+
+
Categorising data in this way makes sense when making choices on how
+to evaluate and analyse data. This is used throughout the
+FreesearchR app to simplify data handling.
Please note that the FreesearchR project is released with a Contributor Code of Conduct. By contributing to this project, you agree to abide by its terms.
+
+
Acknowledgements
+
+
Like any other project, this project was never possible without the great work of others. These are some of the sources and packages I have used: