polmine/fb

Functionality for processing Facebook data (CrowdTangle, Facepager)

fb R Package

Installation

devtools::install_github("PolMine/fb")

Related Work

The online-book „APIs for social scientists: A collaborative review“ includes a review of CrowdTangle access: https://bookdown.org/paul/apis_for_social_scientists/crowdtangle-api.html

The R package RCrowdTangle provides higher-level access to the CrowdTangle API https://github.com/cbpuschmann/RCrowdTangle

fb's Issues

hard-coded removal of special characters in historical_report_purge()

We have relaxed the regex for data cleaning successively. Now we have these steps in historical_report_purge():

  data[["Message"]] <- gsub("\\&", "+", data[["Message"]])
  data[["Message"]] <- gsub("<", "", data[["Message"]])
  data[["Message"]] <- gsub(">", "", data[["Message"]])
  data[["Message"]] <- gsub("[^-0-9a-zA-ZäöüÄÖÜß\\.,!\\?:;#/ ]+", "", data[["Message"]])

It would be better to have a arguments/options rather than this hard-coded approach.

function `historical_report_get_timezone()` and argument 'tz' when reading historical report

The timezones of historical reports may differ, resulting in different values of column "Post Created" for identical posts. When reading the table, the adjustment should consider this, and a helper function historical_report_get_timezone() might assist this by extracting the time zone.

Vignette could explain how to turn facebook data into CWB corpus

Then this snippet of a potential as.xml() function might be useful:

sprintf(
  '<page party="%s" url="%s" date="%s">\n%s\n</page>',
  tbl$`User Name`, tbl$URL, tbl$date, tbl$Message
) %>%
  paste(collapse = "\n") %>%
  sprintf('<?xml version="1.0" encoding="utf-8"?>\n<collection>%s\n</collection>\n', .) %>%
  read_xml()