tutorials/Bernoulli_Example.R

In ctross/STRAND: An R package for simulating and analyzing social network data in R and Stan

########################################
#
#   Bernoulli Analyses  
#
########################################

# Clear working space
rm(list = ls())

# Load libraries
library(STRAND)
library(ggplot2)

#Load package data
data(Colombia_Data)

# Create the STRAND data object
outcome = list(Friends = Colombia_Data$Friends)

dyad = list(Relatedness = Colombia_Data$Relatedness, 
            Distance = Colombia_Data$Distance
            )

groups = data.frame(Ethnicity = as.factor(Colombia_Data$Individual$Ethnicity), 
                    Sex = as.factor(Colombia_Data$Individual$Sex)
                    )

indiv =  data.frame(Age = Colombia_Data$Individual$Age, 
                    BMI = Colombia_Data$Individual$BMI
                     )

rownames(indiv) = rownames(Colombia_Data$Individual)
rownames(groups) = rownames(Colombia_Data$Individual)

dat = make_strand_data(outcome = outcome,
                       block_covariates = groups, 
                       individual_covariates = indiv, 
                       dyadic_covariates = dyad,
                       outcome_mode = "bernoulli",
                       link_mode = "logit")


# Model
fit = fit_block_plus_social_relations_model(data=dat,
                                            block_regression = ~ Sex + Ethnicity,
                                            focal_regression = ~ Age + BMI,
                                            target_regression = ~ Age + BMI,
                                            dyad_regression = ~ Distance + Relatedness,
                                            mode="mcmc",
                                            stan_mcmc_parameters = list(chains = 1, parallel_chains = 1, refresh = 1,
                                                                          iter_warmup = 500, iter_sampling = 500,
                                                                          max_treedepth = NULL, adapt_delta = .98)
)

# Summaries
res = summarize_strand_results(fit)

# Plots
vis_1 = strand_caterpillar_plot(res, submodels=c("Focal effects: Out-degree","Target effects: In-degree","Dyadic effects","Other estimates"), normalized=TRUE, only_slopes=TRUE)
vis_1
#ggsave("Colombia_slopes.pdf", vis_1, width=10, height=5.5)

vis_2 = strand_caterpillar_plot(res, submodels=c("Focal effects: Out-degree","Target effects: In-degree","Dyadic effects","Other estimates"), normalized=FALSE,  only_technicals=TRUE, only_slopes=FALSE)
vis_2
#ggsave("Colombia_corr.pdf", vis_2, width=6, height=2.5)

# Block effects
process_block_parameters(fit, "AFROCOLOMBIAN to AFROCOLOMBIAN", "AFROCOLOMBIAN to EMBERA", HPDI=0.9)
process_block_parameters(fit, "EMBERA to EMBERA", "AFROCOLOMBIAN to EMBERA", HPDI=0.9)
process_block_parameters(fit, "EMBERA to AFROCOLOMBIAN", "AFROCOLOMBIAN to EMBERA", HPDI=0.9)

################################################################# Reciprocity terms
# Simple correlations
cor1 = strand_VPCs(fit, n_partitions = 4, include_reciprocity = TRUE, mode="cor")
cor2 = strand_VPCs(fit, n_partitions = 4, include_reciprocity = TRUE, mode="adj")

df1 = data.frame(cor1[[3]])
colnames(df1) = c("Variable", "Median", "L", "H", "Mean", "SD")
df1$Site = "Unadjusted"
df1$Submodel = rep(c("Generalized","Dyadic"),each=1)

df2 = data.frame(cor2[[3]])
colnames(df2) = c("Variable", "Median", "L", "H", "Mean", "SD")
df2$Site = "Adjusted"
df2$Submodel = rep(c("Generalized","Dyadic"),each=1)

# Adjusted correlations apply only for dyadic reciprocity. Adjusted correlations give the correlation in dyadic random effects+error. Unadjusted give correlation in dyadic random effects.

df = rbind(df1, df2)
df$Median = as.numeric(df$Median)
df$L = as.numeric(df$L)
df$H = as.numeric(df$H)

df$Submodel = factor(df$Submodel)

df$Method = df$Site 

p3 = ggplot2::ggplot(df, ggplot2::aes(x = Variable, y = Median, group = Method, color=Method,
        ymin = L, ymax = H)) + ggplot2::geom_linerange(size = 1,, position = position_dodge(width = 0.6)) + 
        ggplot2::geom_point(size = 2,, position = position_dodge(width = 0.6)) + ggplot2::facet_grid(. ~Submodel, scales = "free", space = "free") +
         ggplot2::geom_hline(ggplot2::aes(yintercept = 0), 
        color = "black", linetype = "dashed") + ggplot2::labs(y = "Regression parameters", 
        x = "") + ggplot2::theme(strip.text.x = ggplot2::element_text(size = 12, 
        face = "bold"), strip.text.y = ggplot2::element_text(size = 12, 
        face = "bold"), axis.text = ggplot2::element_text(size = 12), 
        axis.title.y = ggplot2::element_text(size = 14, face = "bold"), 
        axis.title.x = ggplot2::element_blank()) + ggplot2::theme(strip.text.y = ggplot2::element_text(angle = 360)) + 
        ggplot2::coord_flip() + ggplot2::theme(panel.spacing = grid::unit(1, 
        "lines")) + scale_color_manual(values=c("Unadjusted" = "darkred", "Adjusted" = "black")) + theme(legend.position="bottom")

p3

######################################################################## VPCs
vpc1 = strand_VPCs(fit, n_partitions = 4) # Split variance into focal, target, dyadic, and error componants.
vpc2 = strand_VPCs(fit, n_partitions = 3) # Split variance into focal, target, and dyadic+error componants.

df1 = data.frame(do.call(rbind, vpc1[[2]]))
colnames(df1) = c("Variable", "Median", "L", "H", "Mean", "SD")
df1$Site = "4-way VPC"
df1$Submodel = rep(c("Friendship"),each=4)
df1$Variable2 = rep(c("Focal","Target","Dyadic","Error"),1)

df2 = data.frame(do.call(rbind, vpc2[[2]]))
colnames(df2) = c("Variable", "Median", "L", "H", "Mean", "SD")
df2$Site = "3-way VPC"
df2$Submodel = rep(c("Friendship"),each=3)
df2$Variable2 = rep(c("Focal","Target","Dyadic+Error"),1)

df = rbind(df1, df2)
df$Median = as.numeric(df$Median)
df$L = as.numeric(df$L)
df$H = as.numeric(df$H)

df$Submodel = factor(df$Submodel)
df$Submodel = factor(df$Submodel, levels=c("Friendship"))

df$Variable2 = factor(df$Variable2)
df$Variable2 = factor(df$Variable2, levels=rev(c("Focal","Target","Dyadic","Error","Dyadic+Error")))

df$Type = df$Site 

p4 = ggplot2::ggplot(df, ggplot2::aes(x = Variable2, y = Median, group = Type, color=Type,
        ymin = L, ymax = H)) + ggplot2::geom_linerange(size = 1,, position = position_dodge(width = 0.6)) + 
        ggplot2::geom_point(size = 2,, position = position_dodge(width = 0.6)) + ggplot2::facet_grid(. ~Submodel, scales = "free", space = "free") +
         ggplot2::geom_hline(ggplot2::aes(yintercept = 0), 
        color = "black", linetype = "dashed") + ggplot2::labs(y = "Regression parameters", 
        x = "") + ggplot2::theme(strip.text.x = ggplot2::element_text(size = 12, 
        face = "bold"), strip.text.y = ggplot2::element_text(size = 12, 
        face = "bold"), axis.text = ggplot2::element_text(size = 12), 
        axis.title.y = ggplot2::element_text(size = 14, face = "bold"), 
        axis.title.x = ggplot2::element_blank()) + ggplot2::theme(strip.text.y = ggplot2::element_text(angle = 360)) + 
        ggplot2::coord_flip() + ggplot2::theme(panel.spacing = grid::unit(1, 
        "lines")) + scale_color_manual(values=c("4-way VPC" = "darkred", "3-way VPC" = "black"))  + theme(legend.position="bottom")

p4


###############################################################################################################
############################### Model fit diagnostics
# Note that these functions are run on the raw stan object, so variables are not mapped yet to parameter names.
# See Supplementary Appendix for a list of parameter names
################# Rhat and Effective Samples
# Check all the relevant parameters
fit$fit$summary()
fit$fit$summary("focal_effects")
fit$fit$summary("target_effects")
fit$fit$summary("block_effects")

################# Traceplots
# Check all the relevant parameters
bayesplot::color_scheme_set("mix-blue-red")
bayesplot::mcmc_trace(fit$fit$draws(), pars = c("focal_effects[1]","target_effects[1]","sr_L[2,1]","dr_L[2,1]"))