3 Multidimensional Exploration

In this second section, we simultaneously deal with several dimensions. Multidimensional plotting of variables is then possible.

3.1 Selecting two dimensions with select.dim

When selecting two dimensions (or more) with function select.dim, one ends up with a multidimensional representation of the corresponding variables. For example, in the code below, one selects dimensions week and country (in other words, dimension media is aggregated). Each observation hence gives the number of articles published during a given week and talking about a given country.

geomedia %>%
    select.dim (week, country) %>%
    arrange.elm (country, name) %>%
    arrange.elm (week, name) %>%
    as.data.frame ()

## # A tibble: 11,051 x 3
##    week       country articles
##    <date>     <chr>      <dbl>
##  1 2013-12-30 AFG        22   
##  2 2013-12-30 ARG         6   
##  3 2013-12-30 AUS        29.5 
##  4 2013-12-30 BEN         2   
##  5 2013-12-30 BGD        30   
##  6 2013-12-30 BGR         7.67
##  7 2013-12-30 BHS         2   
##  8 2013-12-30 BLR         0.5 
##  9 2013-12-30 BOL         2.5 
## 10 2013-12-30 BRA        10.5 
## # … with 11,041 more rows

3.2 Plotting bidimensional variables with argument sep.dim.names of plot.var

In order to plot variable articles according to the two selected dimensions, one can use argument sep.dim.names of function plot.var to indicate which one of the two dimensions should be used to separate the data.

geomedia %>%
    select.dim (week, country) %>%
    arrange.elm (week, name) %>%
    filter.elm (country, name %in% c ("USA", "FRA", "DEU")) %>%
    plot.var (articles, sep.dim.names = country, type = "line") +
    theme (axis.text.x = element_text (angle = 90, size = 6))

In this context, using a point-plot (instead of a line-plot) can be useful to separate a dimension with many elements. (Note that, below, we also use filter.elm on week to only plot data for year 2014.)

geomedia %>%
    select.dim (week, country) %>%
    filter.elm (week, format (name, "%Y") == "2014") %>%
    arrange.elm (week, name) %>%
    top_n.elm (country, articles, 20) %>%
    arrange.elm (country, desc (articles)) %>%
    plot.var (articles, sep.dim.names = country, type = "point") +
    theme (axis.text.x = element_text (angle = 90, size = 6))

3.3 Plotting multidimensional variables

It is also possible to separate the data according to several dimensions (with not so many elements) at the same time in order to plot multidimensional variables. For example, the plot below gives the temporal evolution of articles for 8 countries and 2 newspapers.

geomedia %>%
    filter.elm (week, format (name, "%Y") == "2014") %>%
    arrange.elm (week, name) %>%
    top_n.elm (country, articles, 8) %>%
    arrange.elm (country, desc (articles)) %>%
    filter.elm (media, name %in% c ("fr_FRA_lmonde_int", "en_GBR_guardi_int")) %>%
    plot.var (articles, sep.dim.names = list (country, media), type = "point") +
    theme (axis.text.x = element_text (angle = 90, size = 6))