eRum: European R Users Meeting
May 14-16, 2018 @ Budapest, Hungary

What is instant feedback?

Instant Feedback

While collecting declarative data, we would like to provide our respondents feedback relevant to their answer immediately after they answer our question.

Additionaly, we should be able to customise the feedback to each respondent.

The real-world example of this approach could be salary surveys, which ask people about their current salary and compare it to the salaries of people similar to them.

Salary Survey with Instant Feedback
(minimal example)

Demo Screenshoots 1/4

Demo Screenshoots 2/4

Demo Screenshoots 3/4

Demo Screenshoots 4/4

ODGAR Framework

On-line
Declarative Data
Gathering
Analysing
Reporting

ODGAR Framework Design


IQAF Objects provide high level of abstration for declarative data collection with instant feedback.

Each IQAF Object is based on R6Class and can consist of four sub-components:

  • Introduction object (type and content of the introduction to the question)
  • Question object (type and content of the question)
  • Answer object (type and choices of possible answers)
  • Feedback object (type and content of customised feedback)

Each IQAF Object has build-in methods for rendering its compontents and navigating within IQAF object.

Example of simple IQAF object


# create IQAF object with some ID
q1 <- IQAF$new(id = "salary")

# define the question
q1$question <- "How much do you earn?"


# create answer object of type radio as component of our IQAF object
q1$answer <- AnswerRadio$new(id = q1$id)

# define possible choices for our answer
q1$answer$choices <- c("Enough.", "More than I need!", "Less than my neighbour!!")   

The Demo App


ui <- fluidPage(fluidRow(column(width = 12, uiOutput('q1')))) 

server <- function(input, output) {
  output$q1 <- renderUI({
    
    q1$render_iqaf() # rendering the whole IQAF object
    
    }) 
  
  q1$create_observer(input, output) # creating IQAF object observers
  
}

shinyApp(ui = ui, server = server)

ODGAR Goals & Roadmap

The goal is to deliver a useful tool for quickly developing web data products based on declarative data collection with instant feedback, which work with different types of answers and feedbacks, and are mobile ready.

Develop ODGAR R package for

  • creating sets of IQAF objects with dependencies between objects
  • rendering different types of IQAF objects
  • navigating within each IQAF objects and through a sequence of IQAF objects
  • easily implementing different sets of IQAF objects inside a shiny app

Thank you!
Have a great time at eRum2018!

kalimu.github.io/project/odgar