news_search: Search newsfreq.com API

In hrbrmstr/newsfreq: Search for keyword frequencies in news articles via newsfreq.com

Description

news_search provides an interface to the newsfreq.com news keyword search API

Usage

news_search(keywords = NA, target = "", date_from = (Sys.Date() - 1),
  date_to = (Sys.Date() - 1), source = "", summarize = "monthly")

is.newsfreq(x)

Arguments

keywords	search term(s) to query (see Details for specifics)
target	news article component to search in (see Details for valid values and their meaning)
date_from	start date for the search (<= date_to and not current day or future date). Defaults to yesterday. See Details for information on date formatting.
date_to	end date for search (>= date_from and not current day). Defaults to yesterday. See Details for information on date formatting.
source	filter search by news source. You can filter your search by news organization. Entering 'Fox News' or 'The Boston Globe' will search only articles from sources matching that name. Full list available on NewsBank
summarize	Either "monthly" or "annual"

Details

newsfreq.com's interface shows you the frequency that given keywords appear in American News Media and can be used to look at trends in news reporting. news_search provides programmatic access to the search function, returning results in an R data frame where additional analyses can be performed or data can be visualized.

You can use boolean operators AND and OR with parentheses to specify multiple terms in keywords.

The target parameter controls which "field" in news stories are queried. Valid values are:

• "" Search all article fields (the default)

• lead Search in article lead paragraph

• title Search in article headlines

• topics Search in terms identified as index terms by the article provider

• author Search for articles by a particular author

• section Search for articles appearing in a particular news section

• source Search for articles from a particular news source

Search dates must not be in the future and must also not be on the current day (they won't be in the API database). They can be an atomic character vector in a format as.Date will recognize or anything that can be coerced to a Date class.

Value

data.frame (with additional class of newsfreq) of search results. If summarize is "monthly", then the results will also include year, month and month_abb. For "annual", only year will be added to the results.

Note

The "percent" value is a per-annum calculation and will only be calculated and added to the results of searches that are summarized monthly and span full years. It is at best a crude way to normalize the results since the number of news sources (and, hence, articles) changes over time.

Examples

## Not run: 
news_search("data breach", date_from="2014-01-01", date_to="2014-12-31")
##     date_from    date_to year month_abb month count search_date search_terms        pct
## 1  2014-01-01 2014-01-31 2014       Jan    01  3963  2015-02-02  data breach 0.12539155
## 2  2014-02-01 2014-02-28 2014       Feb    02  2856  2015-02-02  data breach 0.09036545
## 3  2014-03-01 2014-03-31 2014       Mar    03  2589  2015-02-02  data breach 0.08191742
## 4  2014-04-01 2014-04-30 2014       Apr    04  2170  2015-02-02  data breach 0.06866002
## 5  2014-05-01 2014-05-31 2014       May    05  2680  2015-02-02  data breach 0.08479671
## 6  2014-06-01 2014-06-30 2014       Jun    06  1973  2015-02-02  data breach 0.06242683
## 7  2014-07-01 2014-07-31 2014       Jul    07  1962  2015-02-02  data breach 0.06207879
## 8  2014-08-01 2014-08-31 2014       Aug    08  2585  2015-02-02  data breach 0.08179086
## 9  2014-09-01 2014-09-30 2014       Sep    09  3326  2015-02-02  data breach 0.10523651
## 10 2014-10-01 2014-10-31 2014       Oct    10  2862  2015-02-02  data breach 0.09055529
## 11 2014-11-01 2014-11-30 2014       Nov    11  2473  2015-02-02  data breach 0.07824711
## 12 2014-12-01 2014-12-31 2014       Dec    12  2166  2015-02-02  data breach 0.06853346

## End(Not run)

hrbrmstr/newsfreq documentation built on May 19, 2019, 8:19 a.m.