In sboysel/afrobarometer: Load Afrobarometer data

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "README-"
)
devtools::install()

afrobarometer: Load Afrobarometer Data into R

The afrobarometer automates laoding Afrobarometer Survey data into the R environment. Open access data is downloaded as needed. The user supplies any optional restricted access data (i.e. survey respondents georeferenced by sampling unit). Merged data is cached in the Afrobarometer data directory and read into R as a local data.frame.

• Data
• Data usage policy

Install

devtools::install_github("sboysel/afrobarometer")

Data

• Open access data: individual responses to survey questionnaires.
• Restricted access data (optional, but must be placed manually): spatial locations, recently administered rounds.

The afrobarometer package fetches any public access data necessary and merges the available spatial data with the questionnaire data. If spatial data is not provided for a specific round, the final table is simply the questionnaire data.

Usage

library(afrobarometer)
afrb_dir(path = "~/Dropbox/dataset/Afrobarometer")

afrb_dir sets the local cache directory to path, creating subdirectories and the local database file as needed. If you have access to spatial location data for each round, place the CSV files in the location subdirectory of the Afrobarometer and name them Locations_R1.csv, Locations_R2.csv, etc. For example, if you have spatial information for rounds 3 and 4, your Afrobarometer data directory should look like this after running afrb_dir:

.
├── locations
│   ├── Locations_R3.csv
│   └── Locations_R4.csv
├── questionnaires
└── codebooks

3 directories, 2 files

Build the database locally

afrb_build(rounds = 3:4, overwrite = TRUE)

Open access questionnaire data and codebooks are downloaded as needed and placed in the Afrobarometer data directory, \code{getOption("afrobarometer.data")}. If spatial data is available, it is merged to questionnaire data for each round. Built datasets are stored in the subdirectory \code{build} of \code{getOption("afrobarometer.data")}. The Afrobarometer directory should now have the following structure:

.
├── build
│   ├── afrb_3.rds
│   └── afrb_4.rds
├── locations
│   ├── Locations_R3.csv
│   └── Locations_R4.csv
├── questionnaires
│   ├── merged_r3_data.sav
│   └── merged_r4_data.sav
└── codebooks
    ├── merged_r3_codebook2_0.pdf
    └── merged_r4_codebook3.pdf

4 directories, 8 files

Pull the merged Round 3 data into R

r3 <- afrb_round(round = 3)

Notes

1. Column names are converted to lowercase on import.
2. As of now, the only variables kepted from the the location files (i.e. Afrobarometer georeferenced data) are the respondent number, latitude, and longitude of the respondent (i.e. respondent's cluster). These variables are then merged with the full questionnaire by respondent number.

Citation

This package simply structures the Afrobarometer data for analysis. Any use of the Afrobarometer data must comply with it's Data Usage and Access Policy. The data is protected by copywright and the authors request the following citation for any use of either the open or restricted data:

Afrobarometer Data, [Country(ies)], [Round(s)], [Year(s)], available at http://www.afrobarometer.org

TODO

• [ ] Package tests (afrobarometer Coverage: 18.28%)
• [x] Download codebooks
• [x] Discuss merging details: take lat + long + respno
• [x] Switch to caching merged data as rds in build directory.
• [ ] Merge multiple rounds together (Inter-round question correspondence, if possible)

sboysel/afrobarometer documentation built on May 29, 2019, 3:24 p.m.