Example Data Analysis

Below is an example of how to analyze the data collected from the Brush Interactions Demo study using R. The data is available in a tidy format, which makes it easy to manipulate and visualize.

1. Install Necessary Packages

list.of.packages <- c("ggplot2", "Hmisc")
new.packages <- list.of.packages[!(list.of.packages %in% installed.packages()[,"Package"])]
if(length(new.packages)) install.packages(new.packages)

library(ggplot2)

2. Read and Preview the Data

df <- read.csv("data/example-brush-interactions_all_tidy.csv")
head(df)

Here is a (truncated) preview of the data:

participantId	trialId	trialOrder	responseId	stage	answer	correctAnswer	duration
b3d13c52	rectangleBrush_q1	3	response	DEFAULT	18	17	16660
b3d13c52	axisBrush_q2	4	max-response	DEFAULT	sun	sun	22369
b3d13c52	axisBrush_q2	4	min-response	DEFAULT	drizzle	drizzle	22369
b3d13c52	sliderBrush_q2	5	max-response	DEFAULT	Gentoo	Gentoo	16142
b3d13c52	sliderBrush_q2	5	min-response	DEFAULT	Chinstrap	Adelie	16142
b3d13c52	rectangleBrush_q2	6	max-response	DEFAULT	Japan	Japan	25142
b3d13c52	rectangleBrush_q2	6	min-response	DEFAULT	Europe	Europe	25142

This is data for one participant. The trialId column indicates the task, answer shows the participant's answer, and correctAnswer shows the correct response, etc. For example, condition sliderBrush_q2 has two responses: a min-response and a max-response.

3. Filter Data for Task q2 and Evaluate Correctness

Now, we want to only look at data for task q2 and check if the participant's answer is correct.

q2 <- subset(df, grepl("_q2", trialId) & status == "completed")



 We can create a new column `isCorrect` that indicates whether the answer matches the correct answer.

```r
q2 <- subset(df, grepl("_q2", trialId) & status == "completed")
q2$isCorrect <- ifelse(q2$answer == q2$correctAnswer, 1, 0)

4. Create a Violin Plot

The plot displays correct answers on the right and incorrect answers on the left.

ggplot(q2, aes(x = isCorrect, y = trialId)) +
  geom_violin(aes(fill = trialId), color = "#888", alpha = 0.7) +
  stat_summary(fun.data = "mean_cl_boot", colour = "#333", size = 0.5, alpha=0.5) +
  theme_minimal() +
  theme(legend.position = "none") +
  labs(
    title = "Violin Plot for q2 (Finding most/least value)",
  )

We find accuracy using paint brush technique is much less than that of the others.

5. Export the generated plot.

ggsave("plot.pdf", width = 5, height = 2, units = "in")