README.md

In pos5737/pos5737data: Data for POS 5737

Datasets for POS 5737

This package contains only datasets—the tidy datasets for POS 5737 found on the data page of the course website.

Quick Start

# install package
devtools::install_github("pos5737/pos5737data")

# see datasets
help(package = "pos5737data")

# load data
data(parties, package = "pos5737data")

Contributing

New datasets require the following:

• Create the R script make-data/get-raw/get-raw-*.R that obtains the raw data (ideally from Dataverse and hopefully versioned) and saves the original format and filename to raw-data/. If raw data are not automatically gettable, then save in raw-data-loc/.
• Create the R script make-data/clean-raw/clean-raw-*.R that tidies the raw data into a easy-to-use data frame and saves it to data/*.rda. Variable names and factor levels should be interpretable. Factor levels should be in a reasonable order. Dates should be store appropriate as date or date-time objects.
• Create a Roxygen2 file R/*-documentation.R that documents the data set. Include brief description of the data set, the dimmensions and descriptions of each variable (it’s okay to point users to the source paper for the details), a link to the raw data files, a link to the published paper(s) using the data, and an example only uses base R functions (and ideally replicates a key result from the original paper).
• Create a test in tests/testthat/test-*.R (usethis::use_test() initiates the file) to test that the data works as expected. Ideally, the test that key results from the paper reproduces.
• Add an example to README.Rmd.

Motivation

In the usual workflow, one might download the data file parties.rds into the data/ directory. They might load the data with

parties <- readr::read_rds("data/parties.rds)

This package allows you to skip the download step.

Usage

Installing the Package

To install the package, run

devtools::install_github("pos5737/pos5737data")

You only need to run this once.

Loading the Data

Once you have the package installed on your computer, you can load the datasets with

# load data
data(parties, package = "pos5737data")

Alternatively, you can use

# load packages
library(pos5737data)

# load data
data("parties")

To see the available datasets, use

help(package = "pos5737data")

Example

parties: Clark and Golder (2006)

# load packages
library(tidyverse)
library(broom)

# load data
data(parties, package = "pos5737data")

# regression model from their table 2, pooled analysis, whole sample
fit <- lm(enep ~ log(average_magnitude)*eneg + upper_tier*eneg + 
            en_pres*proximity, 
          data = parties)

# create table
texreg::screenreg(fit)

## 
## =======================================
##                              Model 1   
## ---------------------------------------
## (Intercept)                    2.81 ***
##                               (0.20)   
## log(average_magnitude)         0.33 ** 
##                               (0.11)   
## eneg                           0.19 *  
##                               (0.08)   
## upper_tier                     0.05 ***
##                               (0.01)   
## en_pres                        0.35 ***
##                               (0.07)   
## proximity                     -3.42 ***
##                               (0.38)   
## log(average_magnitude):eneg    0.08    
##                               (0.06)   
## eneg:upper_tier               -0.02 ***
##                               (0.00)   
## en_pres:proximity              0.80 ***
##                               (0.15)   
## ---------------------------------------
## R^2                            0.30    
## Adj. R^2                       0.29    
## Num. obs.                    555       
## RMSE                           1.59    
## =======================================
## *** p < 0.001, ** p < 0.01, * p < 0.05

therms: ANES Feeling Thermometers

# load data
data(therms)

ggplot(therms, aes(x = ft_democratic_party, y = ft_republican_party)) +
  geom_point()

## Warning: Removed 3059 rows containing missing values (geom_point).

anscombe

data(anscombe, package = "pos5737data")

ggplot(anscombe, aes(x = x, y = y)) + 
  geom_point() + 
  facet_wrap(vars(dataset))

pos5737/pos5737data documentation built on Nov. 5, 2019, 1 a.m.