README.md

In fidadoma/scanpatterns: Set of functions used for working with scan patterns

scanpatterns

R package for analysis of scan patterns

Installation

install.packages("devtools")
devtools::install_github("fidadoma/scanpatterns")

Usage

Let's use example dataset from the package. It is a portion of dataset from Multiple Object Tracking task. First, we need to set global values

library(scanpatterns)
scanpatterns.set.parameters(list("arenamin"=-15,"arenamax"=15,"tmax"=6.0,"spaces-dir"="cache/spaces","eye-dir"="cache/eye"))
data(motData) 
eye1 <- df.eye %>% filter(id == 1, trial == 7) %>% as.eye()
plot(eye1)

Scan pattern comparison

There are several metrics for scan pattern comparison in the package, namely Levenshtein distance Fréchet distance Correlation distance Mean distance * Median distance

We could compute distances between scan patterns. Here we select two scan patterns that were measured in repeated presentation of the same trial.

eye2 <- df.eye %>% filter(id == 1, trial == 13) %>% as.eye()
eye3 <- df.eye %>% filter(id == 1, trial == 34) %>% as.eye()
correlation_distance(eye2, eye3)

Use with your data

Here we show an example how to load data with eyelinker package and convert to objects in this package. See the package description for installation instructions

library(eyelinker)
library(scanpatterns)

ppd <- 60 # we are guessing in this case

fpath <- system.file("extdata/mono500.asc.gz",package="eyelinker")
dat <- read.asc(fpath)
df <- dat$raw %>% 
  group_by(block) %>% 
  mutate(id = 1, trial = block, time = time - min(time) + 1, xp = xp / ppd, yp = yp / ppd) %>% 
  ungroup() %>% 
  select(id, trial, xp, yp, time)

arena <- abs(round(c(range(df$xp), range(df$yp))/ppd*4)/4)

scanpatterns.set.parameters(list("arenamin"=-max(arena),"arenamax"=max(arena),"tmax"=max(df$time)/1000,"max-time"=-1))

df <- df %>% group_by(id, trial) %>% do(eye = as.eye(., xname = "xp", yname = "yp", timename = "time"))
plot(df$eye[[1]])

correlation_distance(df$eye[[1]], df$eye[[2]])

fidadoma/scanpatterns documentation built on May 16, 2019, 12:52 p.m.