In katilingban/ennet: Utilities to Extract and Analyse Text Data from the Emergency Nutrition Network Forum

knitr::opts_chunk$set(
  message = FALSE,
  warning = FALSE,
  error = FALSE,
  collapse = TRUE,
  comment = "#>"
)

library(ennet)
library(dplyr)

The ennet package also includes analytic functions that summarises the text data available from the en-net online forum. Currently, there are four sets of analytic functions available from ennet:

• count_topics_time functions - set of functions to count the number of topics or questions by day, wee, month or year;

• count_topics_theme functions - set of functions to count the number of topics or questions by theme;

• count_topics_author functions - set of functions to count the number of topics or questions by author; and,

• arrange functions - set of functions to arrange topics by number of views or by number of replies.

Counting number of topics/questions

Summarising the number of topics or questions raised within the en-net online forum is basic and useful analytics that can proxy relative importance of a thematic area within the forum. This is facilitated by three sets of functions - 1) count_topics_time; 2) count_topics_theme; and, 3) count_topics_author.

Counting number of topics/questions by time

The count_topics_time set of functions consist of four functions that count topics by day, by week, by month, or by year. To count the number of topics/questions per day, the count_topics_day function is used as follows:

count_topics_day()

which results in:

ennet_topics %>%
  count_topics_day()

By default, the count_topics_day function will provide an output that is not sorted by number of topics/questions per day. To sort by number of topics/questions per day, specify .sort as TRUE:

count_topics_day(.sort = TRUE)

which gives the following output:

ennet_topics %>%
  count_topics_day(.sort = TRUE)

Counting topics/questions by week, month, or year require just using the function named after the time required. For weekly counts, use count_topics_week, for monthly use count_topics_month, and for yearly use count_topics_year.

## Count topics/questions by week
count_topics_week()

## Count topics/questions by month
count_topics_month()

## Count topics/questions by year
count_topics_year()

Counting number of topics/questions by theme

The count_topics_theme set of functions consists of two functions that count topics/questions by theme, and count topics/questions by theme and by time.

Counting of topics by theme is specified as follows:

count_topics_theme()

which results in:

ennet_topics %>%
  count_topics_theme()

Results provided by count_topics_theme is not sorted by number of topics/questions. To sort, specify .sort as TRUE:

count_topics_theme(.sort = TRUE)

which results in:

count_topics_theme(.sort = TRUE)

Counting of topics/questions by theme can also be done by time using the count_topics_theme_time. For example, counting topics/questions by theme by day is done as follows:

count_topics_theme_time(by_time = "day")

which gives the following output:

ennet_topics %>%
  count_topics_theme_time(by_time = "day")

Unlike the previous count functions, the output of count_topics_theme_time is already sorted by default.

To count topics/questions by theme by week, month, or year just requires changing the by_time argument to the desired time interval as shown below:

## Count topics/questions by theme by week
count_topics_theme_time(by_time = "week")

## Count topics/questions by theme by month
count_topics_theme_time(by_time = "month")

## Count topics/questions by theme by year
count_topics_theme_time(by_time = "year")

Counting number of topics/questions by author

The count_topics_author set of functions consists of two functions that count topics/questions by author, and count topics/questions by author and by time.

Counting of topics by author is specified as follows:

count_topics_author()

which results in:

ennet_topics %>%
  count_topics_author()

Results provided by count_topics_author is not sorted by number of topics/questions. To sort, specify .sort as TRUE:

count_topics_author(.sort = TRUE)

which results in:

ennet_topics %>% count_topics_author(.sort = TRUE)

Counting of topics/questions by author can also be done by time using the count_topics_author_time. For example, counting topics/questions by author by day is done as follows:

count_topics_author_time(by_time = "day")

which gives the following output:

ennet_topics %>%
  count_topics_author_time(by_time = "day")

Unlike the previous count functions, the output of count_topics_author_time is already sorted by default.

To count topics/questions by author by week, month, or year just requires changing the by_time argument to the desired time interval as shown below:

## Count topics/questions by author by week
count_topics_author_time(by_time = "week")

## Count topics/questions by author by month
count_topics_author_time(by_time = "month")

## Count topics/questions by author by year
count_topics_author_time(by_time = "year")

Deprecated count functions

In the current release of the ennet package, the following functions have now undergone deprecation:

• count_topics

• count_authors

These functions have now been superseded by the more performant count_topics functions above. We recommend that if you have been using these deprecated functions previously to update your work to use the new functions described above. These deprecated function will be made defunct on the next release of ennet package.

Arranging topics by number of views

Summarising the number of topics or questions raised within the en-net online forum by arranging them based on number of views can proxy level of interest to a specific topic by those participating in the forum. This is facilitated using the arrange_views function. Ranking of topics by number of views is done per thematic area and by a specific time period. Ranking of topics by number of views by thematic area and by month and year is performed by default:

get_themes() %>%
  get_themes_topics() %>%
  arrange_views()

which results in:

ennet_topics %>%
  arrange_views()

Arranging topics by number of views by thematic area and by year is performed as follows:

get_themes() %>%
  get_themes_topics() %>%
  arrange_views(by_date = "year")

which results in:

ennet_topics %>%
  arrange_views(by_date = "year")

Arranging topics by number of views by thematic area overall across the years is performed as follows:

get_themes() %>%
  get_themes_topics() %>%
  arrange_views(by_date = "all")

which results in:

ennet_topics %>%
  arrange_views(by_date = "all")

By default, the output of arrange_views is grouped by thematic area. This default behaviour can be changed by setting the by_theme argument to FALSE. For example, to arrange the topics by number of views by month and year across all themes:

get_themes() %>%
  get_themes_topics() %>%
  arrange_views(by_theme = FALSE)

which results in:

ennet_topics %>%
  arrange_views(by_theme = FALSE)

Arranging topics by number of replies

Summarising the number of topics or questions raised within the en-net online forum by arranging them based on number of replies can proxy level of interest to a specific topic by those participating in the forum specifically those who provide responses and feedback to responses within the discussion. This is facilitated using the arrange_replies function. Ranking of topics by number of replies is done per thematic area and by a specific time period. Ranking of topics by number of replies by thematic area and by month and year is performed by default:

get_themes() %>%
  get_themes_topics() %>%
  arrange_replies()

which results in:

ennet_topics %>%
  arrange_replies()

Arranging topics by number of replies by thematic area and by year is performed as follows:

get_themes() %>%
  get_themes_topics() %>%
  arrange_replies(by_date = "year")

which results in:

ennet_topics %>%
  arrange_replies(by_date = "year")

Arranging topics by number of replies by thematic area overall across the years is performed as follows:

get_themes() %>%
  get_themes_topics() %>%
  arrange_replies(by_date = "all")

which results in:

ennet_topics %>%
  arrange_replies(by_date = "all")

By default, the output of arrange_replies is grouped by thematic area. This default behaviour can be changed by setting the by_theme argument to FALSE. For example, to arrange the topics by number of replies by month and year across all themes:

get_themes() %>%
  get_themes_topics() %>%
  arrange_replies(by_theme = FALSE)

which results in:

ennet_topics %>%
  arrange_replies(by_theme = FALSE)

katilingban/ennet documentation built on Nov. 3, 2022, 4:39 p.m.