Functions for working with {econanalyzr} tibbles
A workflow that showcases
some of the common functions that work with {econanalyzr}
tibbles
Source: vignettes/econanalyzr-verb-workflow.Rmd
econanalyzr-verb-workflow.RmdIntroduction
This vignette walks through a typical econanalyzr
workflow: * Build a valid econanalyzr tibble with the
required schema. * Compute one-off summaries with
econ_value_summary(). * Append a trailing moving average
with econ_calc_trail_avg(). * Filter by a closed date
interval with econ_filter_dates(). * Save the table to a
CSV with a descriptive filename via
econ_csv_write_out().
Construct a valid {econanalyzr} tibble
econanalyzr tables have 9 required columns in a specific order and a
final column viz_type_text. Below we create data for two
geographic entities (US, CA), ensuring doubles/characters as
expected.
set.seed(123)
# Monthly dates for a year
dates <- seq.Date(as.Date("2024-01-01"), by = "month", length.out = 12)
geos <- c("US", "CA")
# Cross-join dates and geos, then create values with simple trends by group
df0 <- tibble::tibble(
date = rep(dates, each = length(geos)),
geo_entity_text = rep(geos, times = length(dates))
) |>
arrange(date, geo_entity_text) |>
group_by(geo_entity_text) |>
mutate(
t_seq = row_number(),
value = as.double(if_else(
geo_entity_text == "US",
100 + 0.20 * t_seq, # gentle uptrend for US
98 + 0.15 * t_seq # gentle uptrend for CA
))
) |>
ungroup() |>
mutate(
date_period_text = "Monthly",
data_element_text = "Quits Rate",
data_measure_text = "Percent",
date_measure_text = "Monthly",
data_transform_text = "Seasonally Adjusted",
geo_entity_type_text = "Nation",
viz_type_text = "Line"
) |>
# Put columns in the required order:
# first 9 required, then viz_type_text last
select(
date, date_period_text, value, data_element_text,
data_measure_text, date_measure_text, data_transform_text,
geo_entity_type_text, geo_entity_text, viz_type_text
)
# Validate & normalize (reorders/coerces if needed)
df <- econanalyzr:::check_econanalyzr_df(df0)
# Peek at the result
df
#> # A tibble: 24 × 10
#> date date_period_text value data_element_text data_measure_text
#> <date> <chr> <dbl> <chr> <chr>
#> 1 2024-01-01 Monthly 98.2 Quits Rate Percent
#> 2 2024-01-01 Monthly 100. Quits Rate Percent
#> 3 2024-02-01 Monthly 98.3 Quits Rate Percent
#> 4 2024-02-01 Monthly 100. Quits Rate Percent
#> 5 2024-03-01 Monthly 98.4 Quits Rate Percent
#> 6 2024-03-01 Monthly 101. Quits Rate Percent
#> 7 2024-04-01 Monthly 98.6 Quits Rate Percent
#> 8 2024-04-01 Monthly 101. Quits Rate Percent
#> 9 2024-05-01 Monthly 98.8 Quits Rate Percent
#> 10 2024-05-01 Monthly 101 Quits Rate Percent
#> # ℹ 14 more rows
#> # ℹ 5 more variables: date_measure_text <chr>, data_transform_text <chr>,
#> # geo_entity_type_text <chr>, geo_entity_text <chr>, viz_type_text <chr>One off summaries with econ_value_summary()
You can compute a statistic on a numeric column after applying either
a date membership filter (dates) or a closed range
(date_range).
# Mean of 'value' for the last 3 months of US only
df_us <- filter(df, geo_entity_text == "US")
last_3 <- tail(unique(df_us$date), 3)
mean_us_last3 <- econ_value_summary(
df = df_us,
dates = last_3,
.fun = mean,
na_rm = TRUE
)
mean_us_last3
#> [1] 102.2
# Median over an *exclusive* range (drop rows in range) for CA
df_ca <- filter(df, geo_entity_text == "CA")
rng <- as.Date(c("2024-04-01","2024-08-01"))
med_ca_exclusive <- econ_value_summary(
df = df_ca,
date_range = rng,
filter_type = "exclusive",
.fun = median
)
med_ca_exclusive
#> [1] 99.35
# Vector summary (quantiles) over all rows
q_all <- econ_value_summary(
df = df_us,
.fun = function(x) stats::quantile(x, probs = c(.25, .5, .75), names = TRUE)
)
q_all
#> 25% 50% 75%
#> 100.75 101.30 101.85Append a trailing moving average in long form
econ_calc_trail_avg() computes a right-aligned trailing
window and appends the derived rows to the original table in long form.
If the input is grouped, the trailing average is computed within
groups.
# Group by geo and compute a 3-period trailing average per series
df_trailing <- df |>
group_by(geo_entity_text) |>
econ_calc_trail_avg(trail_amount = 3L)
# The result contains original rows + trailing-average rows (twice the number of rows here)
nrow(df)
#> [1] 24
nrow(df_trailing)
#> [1] 48
# Show the df_trailing data frame
df_trailing |>
filter(grepl("Trail", data_transform_text))
#> # A tibble: 24 × 10
#> date date_period_text value data_element_text data_measure_text
#> <date> <chr> <dbl> <chr> <chr>
#> 1 2024-01-01 Monthly NA Quits Rate Percent
#> 2 2024-02-01 Monthly NA Quits Rate Percent
#> 3 2024-03-01 Monthly 98.3 Quits Rate Percent
#> 4 2024-04-01 Monthly 98.4 Quits Rate Percent
#> 5 2024-05-01 Monthly 98.6 Quits Rate Percent
#> 6 2024-06-01 Monthly 98.8 Quits Rate Percent
#> 7 2024-07-01 Monthly 98.9 Quits Rate Percent
#> 8 2024-08-01 Monthly 99.0 Quits Rate Percent
#> 9 2024-09-01 Monthly 99.2 Quits Rate Percent
#> 10 2024-10-01 Monthly 99.4 Quits Rate Percent
#> # ℹ 14 more rows
#> # ℹ 5 more variables: date_measure_text <chr>, data_transform_text <chr>,
#> # geo_entity_type_text <chr>, geo_entity_text <chr>, viz_type_text <chr>Filter by a closed date interval
Use econ_filter_dates() to filter by an explicit
interval, or derive the start date from a period
(days/weeks/months/quarters/years). econanalyzr tibbles
will always be in the YYYY-MM-DD format and should be the
earlier date in the period. So the date for Q2 2025 would be
2025-04-01 and for 2026 as a year would be
2026-01-01.
# A) Explicit interval (closed):
filtered_closed <- econ_filter_dates(
df,
start_date = as.Date("2024-06-01"),
end_date = as.Date("2024-10-01")
)
#> Filtered to [2024-06-01, 2024-10-01] (closed interval).
filtered_closed
#> # A tibble: 10 × 10
#> date date_period_text value data_element_text data_measure_text
#> <date> <chr> <dbl> <chr> <chr>
#> 1 2024-10-01 Monthly 99.5 Quits Rate Percent
#> 2 2024-10-01 Monthly 102 Quits Rate Percent
#> 3 2024-09-01 Monthly 99.4 Quits Rate Percent
#> 4 2024-09-01 Monthly 102. Quits Rate Percent
#> 5 2024-08-01 Monthly 99.2 Quits Rate Percent
#> 6 2024-08-01 Monthly 102. Quits Rate Percent
#> 7 2024-07-01 Monthly 99.0 Quits Rate Percent
#> 8 2024-07-01 Monthly 101. Quits Rate Percent
#> 9 2024-06-01 Monthly 98.9 Quits Rate Percent
#> 10 2024-06-01 Monthly 101. Quits Rate Percent
#> # ℹ 5 more variables: date_measure_text <chr>, data_transform_text <chr>,
#> # geo_entity_type_text <chr>, geo_entity_text <chr>, viz_type_text <chr>
# B) Derived interval: "last 6 months" ending at the table’s latest date
filtered_open <- econ_filter_dates(
df,
period_type = "months",
period_amount = 6
)
#> Filtered to [2024-06-01, 2024-12-01] (closed interval).
filtered_open
#> # A tibble: 14 × 10
#> date date_period_text value data_element_text data_measure_text
#> <date> <chr> <dbl> <chr> <chr>
#> 1 2024-12-01 Monthly 99.8 Quits Rate Percent
#> 2 2024-12-01 Monthly 102. Quits Rate Percent
#> 3 2024-11-01 Monthly 99.6 Quits Rate Percent
#> 4 2024-11-01 Monthly 102. Quits Rate Percent
#> 5 2024-10-01 Monthly 99.5 Quits Rate Percent
#> 6 2024-10-01 Monthly 102 Quits Rate Percent
#> 7 2024-09-01 Monthly 99.4 Quits Rate Percent
#> 8 2024-09-01 Monthly 102. Quits Rate Percent
#> 9 2024-08-01 Monthly 99.2 Quits Rate Percent
#> 10 2024-08-01 Monthly 102. Quits Rate Percent
#> # ℹ 4 more rows
#> # ℹ 5 more variables: date_measure_text <chr>, data_transform_text <chr>,
#> # geo_entity_type_text <chr>, geo_entity_text <chr>, viz_type_text <chr>Write to a CSV with a descriptive filename
out_dir <- tempdir()
csv_path <- econ_csv_write_out(
df = filtered_open,
folder = out_dir,
overwrite = TRUE,
quiet = FALSE
)