README.md

In flomuedev/summarize: A collection of helper functions for lazy HCI researchers

lazyhci - helper scripts for lazy HCI researchers

This package is a compilation of functions written over the years for fast and easy exploration of data gathered in scientific experiments. The methods can support you in the areas of plotting and analysis.

WARNING: This package is heavily work in progress. Please trust your common sense more than the results of the package.

Loading

lazy_load can support you in reading and parsing a large amount of csv files into one large table from a folder structure. {*} accepts any characters and ignores them, {VARIABLE}pastes the part of the path as a new variable to the resulting data table.

lazy_load("testdata/logs/day{*}/log_study_{participant}_{*}_{TIME}_{*}_{CONDITION}.csv")

Data Model

All functions in lazy_hci work on a lazy_model of your data. You can create a lazy_model by describing your data:

data.model <- lazy_model(data = data.raw, participant = "SubjectID", within.vars = c("visualization", "pos_undo", "rot_undo"))

The print method provides informations about the model.

Plotting

lazy_plot produces a plot summarizing results for a given dependent variable.

lazy_plot(data.model, dv = "CollectionTime", fun.error = se)

lazy_plot_likert produces a plot using the likert package. lazy_arrange_plots arranges plots in one figure.

All plotting functions support a variety of customization options. Please see the individual documentation. lazyhci returns standard ggplot2 objects which can be further themed and manipulated with the usual methods.

Analysis

lazy_analyze analyzes the given lazy_model using different statistical tools. Currently supported are: aov (all sorts of (RM) ANOVAs fitted using the afex package), art (Aligned-Rank Transformation using the ARTool package), lme,glmer (both using the lme4 package) and friedman (using base-r's friedman.test function. The function automatically checks assumptions and runs corresponding post-hoc tests. Information about assumption checks, the results of the omnibus test and post-hoc tests are availble through the print function. Please see the documentation of the function for more information.

data.analysis.art <- lazy_analyze(lazy_model = data.model, dv = "HappinessRating", analysis_type = "art", posthoc.adj = "bonferroni")

Helper

check_complete_design2 checks for a fully-crossed design for repeated meassures designs. Informs you about missing/NA cells and other problems.

Installing

# install.packages("devtools") ### if you have not installed devtools yet

library(devtools)

install_github("flomuedev/lazyhci")

flomuedev/summarize documentation built on May 22, 2024, 2:23 a.m.