5  Extended example

Typically, package functions include multiple arguments, the different values ​​of which allow for flexible configuration. Below are various examples demonstrating the package’s customization potential.

5.1 Set up

Import the package.

library(paneldesc)

5.2 Data import

Import the built-in dataset with simulated unbalanced panel data.

data(production)

Set up a panel structure in advance so that you don’t have to do it later each time you use other functions. Note that if delta is supplied, the function checks for omitted time periods. If such periods exist, they will be taken into account when other functions work with this argument.

panel <- make_panel(production, index = c("firm", "year"), delta = 1)

5.3 Panel data structure analysis

If necessary, you can display more detailed statistics on the distribution of entities by periods and periods by entities.

describe_balance(panel, detail = TRUE, digits = 2)
dimension mean std min p5 p25 p50 p75 p95 max
entities 26.17 3.97 19 20.5 25.25 27 28.75 29.75 30
periods 5.23 0.94 3 4.0 4.25 6 6.00 6.00 6

plot_periods() function allows to customize colors.

plot_periods(panel, colors = c("gray", "black"))

Both describe_patterns() and plot_patterns() allow you to limit the range of patterns to display. In addition, desribe_patterns() allows to customize rounding when calculting shares, while plot_patterns() allows to customize colors.

describe_patterns(panel, limits = 3, digits = 2)
pattern 1 2 3 4 5 6 count share
1 1 1 1 1 1 1 16 0.53
2 1 1 1 1 1 0 5 0.17
3 1 1 1 1 0 0 3 0.10 
plot_patterns(panel, limits = c(4, 6), colors = c("darkgray", "white"))

5.4 Missing values analysis

plot_missing() allows to select specific variables to analyze. One can also customize colors.

plot_missing(
  panel,
  select = c("labor", "capital"),
  colors = c("black", "white")
)

summarize_missing() also allows to select specific variables to analyze. In addition, it can provide more detailed table by adding period-specific missing value counts. One can also customize rounding.

summarize_missing(
  panel,
  select = c("labor", "capital"),
  detail = TRUE,
  digits = 2
)
variable na_count na_share entities periods 1 2 3 4 5 6
labor 26 0.14 15 6 6 2 1 1 5 11
capital 26 0.14 17 6 5 2 2 2 4 11

describe_incomplete() can provide more detailed table by adding variable-specific missing value counts.

describe_incomplete(panel, detail = TRUE)
firm na_count variables sales capital labor industry ownership region
23 18 6 3 3 3 3 3 3
6 13 6 2 2 3 2 2 2
7 13 6 3 2 2 2 2 2
1 12 6 2 2 2 2 2 2
2 12 6 2 2 2 2 2 2
12 12 6 2 2 2 2 2 2
21 12 6 2 2 2 2 2 2
26 12 6 2 2 2 2 2 2
25 7 6 1 1 2 1 1 1
30 7 6 2 1 1 1 1 1
4 6 6 1 1 1 1 1 1
13 6 6 1 1 1 1 1 1
17 6 6 1 1 1 1 1 1
29 6 6 1 1 1 1 1 1
14 2 2 1 0 1 0 0 0
10 1 1 0 1 0 0 0 0
22 1 1 0 1 0 0 0 0
27 1 1 0 1 0 0 0 0

5.5 Numeric variables analysis

summarize_numeric() allows to select specific variables to analyze. It also can provide additional statisitcs. Custom rounding is also available.

summarize_numeric(
  panel,
  select = c("sales", "labor"),
  group = "year",
  detail = TRUE,
  digits = 2
)
year variable count mean std min p25 p50 p75 max
1 sales 25 72.30 43.04 12.00 44.77 70.21 89.97 192.59
1 labor 24 74.35 69.77 18.41 32.64 58.58 76.74 333.38
2 sales 28 67.56 35.06 20.37 40.72 64.69 90.57 150.74
2 labor 28 81.76 66.57 17.67 37.43 66.14 89.84 327.58
3 sales 29 72.09 64.52 17.32 33.49 46.72 80.69 292.85
3 labor 29 82.26 106.87 8.17 31.73 57.19 91.43 579.02
4 sales 27 56.86 31.71 12.00 36.25 48.24 67.06 136.69
4 labor 29 55.53 46.45 5.97 26.86 40.51 61.74 207.57
5 sales 26 62.16 41.24 19.77 33.57 49.30 69.85 189.07
5 labor 25 74.06 71.26 15.88 32.17 49.53 87.45 266.83
6 sales 19 83.83 45.56 25.99 46.48 78.38 115.12 191.59
6 labor 19 101.00 67.26 18.67 48.59 79.96 130.89 259.85

plot_heterogeneity() allows to choose several grouping variables. One can also customize colors.

plot_heterogeneity(
  panel,
  select = "sales",
  group = c("firm", "industry", "year"),
  colors = c("black", "gray")
)

decompose_numeric() allows to select specific varibles to analyze. One also can output the resulting table in wide format and hide redundant details, as well as customize rounding.

decompose_numeric(
  panel,
  select = c("sales", "labor"),
  detail = FALSE,
  format = "wide",
  digits = 2
)
variable mean std_overall std_between std_within
sales 68.40 45.02 29.06 34.13
labor 76.88 74.15 41.07 61.20

plot_demeaned() allows to choose several grouping variables. One can also customize colors.

plot_demeaned(
  production,
  select = "sales",
  group = c("industry", "region", "year"),
  colors = c("gray", "black")
)

plot_demeaned(
  production,
  select = c("labor", "sales"),
  group = c("industry", "region", "year"),
  colors = c("gray", "black")
)

5.6 Factor variables analysis

Compared to the default settings, both decompose_factor() and summarize_transition() functions can output the resulting table in long format and also allow to customize rounding. descompose_factor() also allows to select specific variables to analyze.

decompose_factor(panel, select = "industry", format = "long", digits = 2)
variable category dimension count share
industry Industry 1 between 13 0.43
industry Industry 2 between 11 0.37
industry Industry 3 between 10 0.33
industry Industry 1 overall 63 0.40
industry Industry 2 overall 45 0.29
industry Industry 3 overall 49 0.31
industry Industry 1 within NA 0.92
industry Industry 2 within NA 0.81
industry Industry 3 within NA 0.92 
summarize_transition(panel, select = "industry", format = "long", digits = 2)
23 rows with NA values in 'industry' removed.
from to count share
Industry 1 Industry 1 50 1.00
Industry 1 Industry 2 0 0.00
Industry 1 Industry 3 0 0.00
Industry 2 Industry 1 2 0.05
Industry 2 Industry 2 34 0.92
Industry 2 Industry 3 1 0.03
Industry 3 Industry 1 0 0.00
Industry 3 Industry 2 1 0.03
Industry 3 Industry 3 39 0.98