djhocking

# make fake dataset
df <- data.frame(x = runif(100, 0, 1), y = rnorm(100, 10, 3), z = rpois(100, 10))

# subset dataframe
df_sub <- df[which(df$x >= 0.75), ]

# subset using dplyr
library(dplyr)

df_sub2 <- df %>%

djhocking

Predict vs simulate in lme4

For this investigation we are going to use the sleepdata data set from the lme4 package. Here is the head of the data frame:

djhocking

# ggplot boxplot groups of continuous x

### Daniel J. Hocking

I am trying to make a boxplot with ggplot2 in R where the x-axis in continous but their are paired boxplots for each value of on the x-axis based on another factor with two possible values. I want to make a plot where boxplots are arranged by number of survey years on the x-axis but paired by spatialTF (2 boxplots for every value of n_years) but n_years are not evenly spaced.

This plot gets the paired boxplots correct by year but the years on the x-axis are evenly spaced and don't reflect the actual (continuous) time between years.

```
ggplot(df_converged, aes(factor(n_years), mean_N_est)) + 

djhocking

# table references
tbl_locations <- tbl(db, 'locations') %>%
rename(location_id=id, location_name=name, location_description=description) %>%
select(-created_at, -updated_at)
tbl_series <- tbl(db, 'series') %>%
rename(series_id=id) %>%
select(-created_at, -updated_at)
tbl_variables <- tbl(db, 'variables') %>%
rename(variable_id=id, variable_name=name, variable_description=description) %>%

djhocking

# Get list of unique catchments with daymet data in our database
drv <- dbDriver("PostgreSQL")
con <- dbConnect(drv, dbname=...)
# con <- dbConnect(drv, dbname=...)
qry <- "SELECT DISTINCT featureid FROM daymet;"
result <- dbSendQuery(con, qry)
catchments <- fetch(result, n=-1)
catchments <- as.character(catchments$featureid)
 
# get daymet data for a subset of catchments

djhocking

# fetch temperature data
tbl_values <- left_join(tbl_series,
                        select(tbl_variables, variable_id, variable_name),
                        by=c('variable_id'='variable_id')) %>%
  select(-file_id) %>%
  filter(location_id %in% df_locations$location_id,
         variable_name=="TEMP") %>%
  left_join(tbl_values,
            by=c('series_id'='series_id')) %>%
  left_join(select(tbl_locations, location_id, location_name, latitude, longitude, featureid=catchment_id),

djhocking

We assumed stream temperature measurements were normally distributed following,

\\[ t_{s,h,d,y} \sim \mathcal{N}(\mu_{s,h,d,y}, \sigma) \\]

where $t_{s,h,d,y}$ is the observed stream water temperature at the site ($s$) within the sub-basin identified by the 8-digit Hydrologic Unit Code (HUC8; $h$) for each day ($d$) in each year ($y$). We describe the normal distribution with the standard deviation ($\sigma$). The expected temperature follows a linear trend

\\[ \omega_{s,h,d,y} = X^0 B^0 + X_{h}^{huc} B_{h}^{huc} + X_{s,h}^{site} B_{s,h}^{site} + X_{y}^{year} B_{y}^{year} \\]

but the expected temperature ($\mu_{s,h,d,y}$) is adjusted based on the residual error from the previous day

djhocking

#purling is in the knitr package
library(knitr)


setwd("C:/ALR/Models/boo") #example using local windows directory, can easily switch

#it can be really simple
purl( "script1.Rmd", "script1.R" )
#just specify the rmd filename, then r filename, with extensions

djhocking

one <- seq(1:10)
two <- rnorm(10)
three <- runif(10, 1, 2)
four <- -10:-1

df <- data.frame(one, two, three)
df2 <- data.frame(one, two, three, four)

str(df)

djhocking

# Function for getting bootstrapped glmer predictions in parallel
glmmBoot <- function(dat, form, R, nc){
  # dat = data for glmer (lme4) logistic regression
  # form = formula of glmer equation for fitting
  # R = total number of bootstrap draws - should be multiple of nc b/c divided among cores evenly
  # nc = number of cores to use in parallel

  library(parallel)
  cl <- makeCluster(nc) # Request # cores
  clusterExport(cl, c("dat", "form", "nc", "R"), envir = environment()) # Make these available to each core