7.1 Examples
A few examples of DGP.R are provided below.
Example 2
Description: Assign the full masterdata to each student (no sampling). This results in all students receiving the same dataset (adapted from Homework 4).
Example 3
Description: Randomly assign each student 20,000 observations sampled from every masterdata except for the masterdata corresponding to datasetB.csv.6
For the masterdata corresponding to datasetB.csv, assign each student the full masterdata. Assume that datasetB.csv has exactly two columns, col_1 and col_2 (adapted from Homework 7).
DGP <- function(masterdata = "Masterdata.csv", PERMID = PERMID, dataname = dataname){
masterdata <- read_csv(masterdata)
set.seed(PERMID)
#---------------#
#Edit made here...
#No sampling for datasetB.csv...
if(all(colnames(masterdata) %in% c("col_1", "col_2"))){
d <- masterdata
#Everything else is randomly sampled...
} else{
d <- masterdata[sample(1:nrow(masterdata), 20000, replace = F),]
}
#---------------#
readr::write_csv(d, dataname)
}Example 4
Description: For the following masterdata, assign as follows:
data1.csv: Randomly sample 10,000 rows.data2.csv: Randomly sample 20,000 rows.data3.csv: Assign the full dataset, arrange by its columndata3_col2(so every student will get their dataset in this format).data4.csv: Assign the full dataset.
Assume data1.csv has some unique column data1_unique_column and that data2_col1, … ,data2_col5 are the only columns in data2.csv and data3_col1 and data3_col2 are the only columns in data3.csv (adapted from Homework 14).
DGP <- function(masterdata = "Masterdata.csv", PERMID = PERMID, dataname = dataname){
masterdata <- read_csv(masterdata)
set.seed(PERMID)
#---------------#
#Edit made here...
#`data1.csv` randomly sampled for 10,000 observations
if("data1_unique_column" %in% colnames(masterdata)){
d <- masterdata[sample(1:nrow(masterdata), 10000, replace = F),]
#`data2.csv` randomly sampled for 20,000 observations
} else if(all(colnames(masterdata) %in% c("data2_col1", "data2_col2",
"data2_col3", "data2_col4", "data2_col5"))){
d <- masterdata[sample(1:nrow(masterdata), 20000, replace = F),]
#`data3.csv` is not sampled and is arranged by `data3_col2`
} else if(all(colnames(masterdata) %in% c("data3_col1", "data3_col2"))){
d <- masterdata %>% arrange(data3_col2)
#`data4.csv` is not sampled
} else{
d <- masterdata[sample(1:nrow(masterdata), nrow(masterdata), replace = F),]
}
#---------------#
readr::write_csv(d, dataname)
}Example 5
Description: Assign each student a dataset containing observations from 140 randomly selected countries from masterdata’s Country or Area column (adapted from Homework 8).
DGP <- function(masterdata = "Masterdata.csv", PERMID = PERMID, dataname = dataname){
masterdata <- read_csv(masterdata)
set.seed(PERMID)
#---------------#
#Edit made here...
#Creating a list of all the countries in `masterdata`
country_list <- unique(masterdata$`Country or Area`)
#Randomly selecting 140 countries
country_sample <- sample(country_list, 140)
#Selecting the rows from `masterdata` that contain the selected countries
d <- masterdata[masterdata$`Country or Area` %in% country_sample, ]
#---------------#
readr::write_csv(d, dataname)
}Sometimes, multiple csv files are required for an assignment, creating multiple
masterdata. TheDGP.Rfile will then loop through eachmasterdata.↩︎