other_R_code/old vignettes that need cleaning/my-vignette.utf8.md

In gidonc/durhamevp: ESRC/AHRC Victorian Election Violence Project R Code (Durham University)

title: "Interacting with the Durham Election Violence Database Using R" author: "Gidon Cohen" date: "2018-07-24" output: rmarkdown::html_vignette vignette: > %\VignetteIndexEntry{Interacting with the Durham Election Violence Database Using R} %\VignetteEngine{knitr::rmarkdown} %\VignetteEncoding{UTF-8}

Introduction

This vignette documents how to interact with the Durham election violence database using R. To use the package download it from the local repository. You can then load all the functions with the following command:

library(durhamevp)

## also using tidyverse functions
library(tidyverse)

Connecting to the Database

To connect to the database use the evdb_connect function. The function needs to pass the password to the database. The default option (password_method = "ask") will prompt you for the password. This "ask" method for obtaining the password only works in RStudio and only in interactive mode (e.g. not when writing Rnw or Rmd files).

con <- evdb_connect()

Once you have given the connection a name (here con) you will pass this as a parameter to the other database function.

To pass the password without interactive modes use either the "config" package or the "keyring" package ("keyring" is more secure and should be the first option). These methods might be more useful long-term as the password only has to be entered once per computer, and these methods work without user interaction (e.g. when using knitr).

con <- evdb_connect(password_method="keyring")

Allocating Documents to Users

Allocation of the intial document sets (training and test sets)

Most new users will need both the training set and the test set allocated to them. The function assign_initialsets_to_user performs this function for a single user or for groups of users. The first argument is the database connection and the second is the user id or user ids. Thus, to allocate the testset and the training set to user 3:

assign_initalsets_to_users(3)
#> Warning: package 'bindrcpp' was built under R version 3.4.4
#> assign_testset_to_user allocated 0 items to user 3. 20 items (of 20) were already allocated
#> assign_trainingset_to_user allocated 0 items to user 3. 16 items (of 16) were already allocated

Or to allocate the testset and training set to user 3, 4, 5 and 6:

assign_initalsets_to_users(c(3, 4, 5, 6))
#> assign_testset_to_user allocated 0 items to user 3. 20 items (of 20) were already allocated
#> assign_trainingset_to_user allocated 0 items to user 3. 16 items (of 16) were already allocated
#> assign_testset_to_user allocated 0 items to user 4. 20 items (of 20) were already allocated
#> assign_trainingset_to_user allocated 0 items to user 4. 16 items (of 16) were already allocated
#> assign_testset_to_user allocated 0 items to user 5. 20 items (of 20) were already allocated
#> assign_trainingset_to_user allocated 0 items to user 5. 16 items (of 16) were already allocated
#> assign_testset_to_user allocated 0 items to user 6. 20 items (of 20) were already allocated
#> assign_trainingset_to_user allocated 0 items to user 6. 16 items (of 16) were already allocated

Note that if a user already has an item from the testset or the training set allocated to them this function does not allocate the article again. Thus, in the above example all the users already had the full training and test sets allocated to them so running the function did not alter the data.

The function will allocate the training set as type training and the test set as type testing, but it does not alter the type of existing alloctions (so if an item from the testset was allocated as type training this would not be altered).

The function assign_trainingset_to_user and assign_testset_to_user make it easy to assign these parts of the initial document sets to users if that is desired instead (currently these functions are configured to take just a single user_id at a time).

Related Database Functions

View Allocations

To see the whole set of allocations belonging to a user use the get_allocation function. The function has an allocation_type parameter which can take the values training, testing, coding, checking and ideal (where ideal is for the gold standard coding) which restrict the results to allocations of that type. The default allocation type is "all" which returns all allocation_types. Thus, to view the whole allocation for user 1 (the select command just reduces the columns for display:

get_allocation(user_id = 1)%>%
  dplyr::select(user_id, document_id, allocation_type, allocation_date) 
#>   user_id document_id allocation_type allocation_date
#> 1       1         265        training      2018-02-24
#> 2       1           1        training      2018-02-24
#> 3       1           1     ideal-debug      2018-04-16
#> 4       1           1        training      2018-04-17
#> 5       1           3        training      2018-04-25
#> 6       1           1         testing      2018-05-23

To view their allocation where the allocation type is ideal:

get_allocation(user_id = 1, allocation_type="ideal") %>%
  dplyr::select(user_id, document_id, allocation_type, allocation_date) 
#> [1] user_id         document_id     allocation_type allocation_date
#> <0 rows> (or 0-length row.names)

Check Allocations

To check if a user has the training set allocated to them

gidonc/durhamevp documentation built on April 8, 2022, 10:31 a.m.