| Type: | Package |
| Title: | Spaghetti-Plot Fixed and Random Effects of Linear Mixed Models |
| Version: | 0.1.0 |
| Description: | Plot both fixed and random effects of linear mixed models, multilevel models in a single spaghetti plot. The package allows to visualize the effect of a predictor on a criterion between different levels of a grouping variable. Additionally, confidence intervals can be displayed for fixed effects. Calculation of predicted values of random effects allows only models with one random intercept and/or one random slope to be plotted. Confidence intervals and predicted values of fixed effects are computed using the 'ggpredict' function from the 'ggeffects' package. Lüdecke, D. (2018) <doi:10.21105/joss.00638>. |
| Encoding: | UTF-8 |
| LazyData: | true |
| Imports: | ggeffects, ggplot2, glmmTMB, lme4 |
| RoxygenNote: | 7.3.1 |
| Depends: | R (≥ 2.10) |
| License: | MIT + file LICENSE |
| NeedsCompilation: | no |
| Packaged: | 2024-05-17 10:48:22 UTC; jfp |
| Author: | Jan-Felix Palnau 
[aut, cre] |
| Maintainer: | Jan-Felix Palnau <jan.palnau@mailbox.org> |
| Repository: | CRAN |
| Date/Publication: | 2024-06-05 19:30:08 UTC |
Spaghetti-plot fixed and random effects of linear mixed models
Description
cookPasta() creates dataframes from fixed and random effects of 'lme4' or 'glmmTMB' models (e.g., for plotting)
Usage
cookPasta(
model = NULL,
predictor = NULL,
nested.in = NULL,
group = NULL,
ci.int = FALSE,
ci.lvl = ci.lvl
)
Arguments
model |
lme4 or glmmTMB model object |
predictor |
(Character) Name of predictor (e.g., "time" or "math_score"), as it is present in the model |
nested.in |
(Character) Name of the variable your time points or subjects are nested in (e.g.,"school" or "id") |
group |
(Optional, character) The name of your grouping variable (e.g., "condition" or "gender") |
ci.int |
(Optional, boolean) Enable confidence (prediction) intervals, disabled by default |
ci.lvl |
(Optional, numeric) Set level of confidence (prediction) intervals (default: 0.95). Requires ci.int to be set to TRUE |
Value
Returns a list of two dataframes, in which the first element is the fixed effects dataframe and the second element the random effects dataframe
ECOVIA data
Description
A subset of data from the ECOVIA app
Usage
ecovia_data
Format
A data frame with 1743 rows and 4 columns:
- id
Subjects, in which time points are nested in
- time
Day of intervention
- condition
Condition (control vs. intervention)
- CO2
Daily dietary carbon emissions
Source
<https://osf.io/qd7vw/?view_only=cc22e0d1de8844e1850d8ef1442fbecb>
JSP data
Description
A subset of JSP secondary school exam data
Usage
jsp_data
Format
A data frame with 728 rows and 7 columns:
- math_score_y1
Score on math test year 1
- math_score_y3
Score on math test year 3
- gender
Gender of subjects (pupils)
- social_class
Social class of subjects (pupils)
- school
School, in which subjects (pupils) are nested in
- normal_score_y1
Normal test score year 1
- normal_score_y3
Normal test score year 3
Source
<https://www.bristol.ac.uk/cmm/team/hg/msm-3rd-ed/datasets.html>
Spaghetti-plot fixed and random effects of linear mixed models
Description
pastaPlot() plots slopes for both fixed and random effects of linear mixed models from 'lme4' or 'glmmTMB' packages as a single spaghetti plot, optionally between conditions including confidence bands for fixed effects.
Usage
pastaPlot(
model = NULL,
predictor = NULL,
nested.in = NULL,
group = NULL,
legend.title = "Legend",
group.labels = NULL,
xlab = NULL,
ylab = NULL,
font.family = NULL,
colors = NULL,
ci.lvl = 0.95,
ci.int = FALSE,
ci.linetype = 0,
lwd.fix = 1,
lwd.ran = 0.5,
xlab.inc = 0,
xlab.int = NULL,
ylim = NULL,
opacity.ci = 0.25,
opacity.ran = 0.3,
colors.ci = NULL
)
Arguments
model |
lme4 or glmmTMB model object |
predictor |
(Character) Name of predictor (e.g., "time" or "math_score"), as it is present in the model |
nested.in |
(Character) Name of the variable your time points or subjects are nested in (e.g.,"school" or "id") |
group |
(Optional, character) The name of your grouping variable (e.g., "condition" or "gender") |
legend.title |
(Optional, character) Name of legend in plot (e.g., "Condition", or "Gender") |
group.labels |
(Optional, vector of characters) Names of group labels to be displayed in the plot (e.g., c("Control", "Intervention")) |
xlab |
(Optional, character) Label of x-axis (predictor) (e.g., "Time (days)") |
ylab |
(Optional, character) Label of y-axis (dependant variable) (e.g., "GAF") |
font.family |
(Optional, character) Name of the font family (e.g. "serif") |
colors |
(Optional, vector of characters) Set color of slopes. Length of vector should correspond to number of values in group variable (e.g., c("#5e9aff", "blue")). If no group variable is specified, pass a single color. |
ci.lvl |
(Optional, numeric) Set confidence interval (default: 0.95) |
ci.int |
(Optional, logical) Enable confidence (prediction) intervals, disabled by default |
ci.linetype |
(Optional, numeric) Set linetype of confidence bands outline (default: 0) |
lwd.fix |
(Optional, numeric) Line width of fixed effects (default: 1) |
lwd.ran |
(Optional, numeric) Line width of random effects (default: 0.5) |
xlab.inc |
(Optional, numeric) Increment the displayed values of your predictor (e.g., xlab_int = 1 changes range of x from 0-29 to 1-30), set to 0 by default |
xlab.int |
(Optional, numeric) Interval between displayed predictor values on x-axis (e.g., "1"), disabled by default |
ylim |
(Optional, numeric vector) Limited range of values on y-axis (e.g. c(1,5.5)) |
opacity.ci |
(Optional, numeric) Set opacity of confidence bands in the range of 0 to 1 (default = 0.1) |
opacity.ran |
(Optional, numeric) Set opacity of random slopes in the range of 0 to 1 (default = 0.4) |
colors.ci |
(Optional, vector of characters) Set color of confidence bands. Length of vector should correspond to number of values in group variable (e.g., c("#5e9aff", "blue")). If no group variable is specified, pass a single color. |
Value
Returns a ggplot2 plot object to further be modified
Examples
lme4_model <- lme4::lmer(CO2 ~ 1 + time*condition + (1 + time | id),
data=ecovia_data, REML = FALSE, control = lme4::lmerControl(optimizer = "bobyqa"))
pastaPlot(lme4_model, "time", "id", group = "condition", legend.title = "Condition",
group.labels = c("Control", "Intervention"), ci.int = TRUE, xlab = "Time (days)",
ylab = "CO2")
glmmTMB_model <- glmmTMB::glmmTMB(math_score_y3 ~ 1 + math_score_y1*gender +
(1 + math_score_y1 | school), data=jsp_data, REML = FALSE)
pastaPlot(glmmTMB_model, "math_score_y1", "school", group = "gender",
legend.title = "Gender", group.labels = c("Male", "Female"), ci.int = FALSE,
xlab = "Math score (year 1)", ylab = "Math score (year 3)")