Compute Percentage Changes of School Resources and Population

library(tibble)
library(dplyr)
#> 
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#> 
#>     filter, lag
#> The following objects are masked from 'package:base':
#> 
#>     intersect, setdiff, setequal, union
library(tidyr)
library(readr)
library(kableExtra)
#> 
#> Attaching package: 'kableExtra'
#> The following object is masked from 'package:dplyr':
#> 
#>     group_rows

library(PrjCompPPTS)

# If resave outputs to data, only do this during development
bl_resave_to_data <- FALSE

In this file, we compute percentage changes in school resources and population. We interpolate when there is gaps in the data.

For all variables, for each country/location, we compute:

• percentage changes year by year
• percentage changes every 5, 10, 15, and 20 years.

We interpolate and generate interval changes for all countries. We also extrapolate for up to 3 years, to reach the closest decade break-off points.

To illustrate what we do, we print results from Afghanistan and Austria, where there are gaps in the data.

In the case of Afghanistan, we:

1. Interpolate to fill-in gap years, 1983, 1987, and 1992, for example.
2. Extrapolate to fill-in value for 2020 given 2018 to 2019 changes.

In the case of Austria, in earlier decades, data is not annual, so we interpolate to obtain more annualized predictions.

We also show results illustration from Busan Korea.

Load and select

# A. Load and select ----
ppts_easia_weuro_sel <- ppts_easia_weuro_world %>%
  select(location_code, location_level, year,
         contains("stats"), -stats_enroll_ratio)

Transform data

# B. Long to wide ----
ppts_easia_weuro_long <- ppts_easia_weuro_sel %>%
  pivot_longer(cols = starts_with('stats'),
               names_to = c('variable'),
               names_pattern = paste0("stats_(.*)"),
               values_to = "value")
str(ppts_easia_weuro_long)
#> tibble [88,560 × 5] (S3: tbl_df/tbl/data.frame)
#>  $ location_code : Factor w/ 286 levels "ABW","AFE","AFG",..: 1 1 1 1 1 1 1 1 1 1 ...
#>  $ location_level: Factor w/ 4 levels "country","multicountry",..: 1 1 1 1 1 1 1 1 1 1 ...
#>  $ year          : num [1:88560] 1960 1960 1960 1960 1960 ...
#>  $ variable      : chr [1:88560] "youthpop" "student" "teacher" "school" ...
#>  $ value         : num [1:88560] 23769 NA NA NA NA ...
# View(ppts_easia_weuro_long)

# C sort and group ----
ppts_easia_weuro_long <- ppts_easia_weuro_long %>%
  arrange(location_code, location_level, variable, year) %>%
  group_by(location_code, location_level, variable)
# kable(ppts_easia_weuro_long %>% 
#   filter(location_code == 'AFG' & variable == 'student'))

Annual percentage changes, Interpolate and Extrapolate

Raw percentage changes

We compute annual percentage changes below. Note that

# D. Annual percentage changes ----
# Compute these
# - annual: for all possible consecutive years
# - annual_interp1: annual based on consecutive year if possible, 
# when that is not possible, find closest years of available data, 
# and derive annual (considering compounding) growth rates

# D.1 annual
ppts_easia_weuro_long <- ppts_easia_weuro_long %>%
  mutate(pchg_yr1 = (value - lag(value)) / lag(value))

# View(ppts_easia_weuro_long)
kable(ppts_easia_weuro_long %>% 
        filter(location_code == 'AFG' & variable == 'student'))

location_code	location_level	year	variable	value	pchg_yr1
AFG	country	1960	student	NA	NA
AFG	country	1961	student	NA	NA
AFG	country	1962	student	NA	NA
AFG	country	1963	student	NA	NA
AFG	country	1964	student	NA	NA
AFG	country	1965	student	NA	NA
AFG	country	1966	student	NA	NA
AFG	country	1967	student	NA	NA
AFG	country	1968	student	NA	NA
AFG	country	1969	student	NA	NA
AFG	country	1970	student	540685	NA
AFG	country	1971	student	572933	0.0596429
AFG	country	1972	student	597509	0.0428951
AFG	country	1973	student	624374	0.0449617
AFG	country	1974	student	654209	0.0477839
AFG	country	1975	student	692342	0.0582887
AFG	country	1976	student	729667	0.0539112
AFG	country	1977	student	764175	0.0472928
AFG	country	1978	student	810657	0.0608264
AFG	country	1979	student	NA	NA
AFG	country	1980	student	959583	NA
AFG	country	1981	student	1024741	0.0679024
AFG	country	1982	student	365458	-0.6433655
AFG	country	1983	student	NA	NA
AFG	country	1984	student	447351	NA
AFG	country	1985	student	479150	0.0710829
AFG	country	1986	student	512809	0.0702473
AFG	country	1987	student	NA	NA
AFG	country	1988	student	651622	NA
AFG	country	1989	student	622135	-0.0452517
AFG	country	1990	student	622513	0.0006076
AFG	country	1991	student	627888	0.0086344
AFG	country	1992	student	NA	NA
AFG	country	1993	student	786532	NA
AFG	country	1994	student	1161444	0.4766646
AFG	country	1995	student	1312197	0.1297979
AFG	country	1996	student	NA	NA
AFG	country	1997	student	NA	NA
AFG	country	1998	student	1046338	NA
AFG	country	1999	student	875605	-0.1631719
AFG	country	2000	student	749360	-0.1441803
AFG	country	2001	student	773623	0.0323783
AFG	country	2002	student	2667629	2.4482287
AFG	country	2003	student	3781015	0.4173691
AFG	country	2004	student	4430142	0.1716806
AFG	country	2005	student	4318819	-0.0251285
AFG	country	2006	student	4669110	0.0811081
AFG	country	2007	student	4718077	0.0104874
AFG	country	2008	student	4974836	0.0544203
AFG	country	2009	student	4945632	-0.0058703
AFG	country	2010	student	5279326	0.0674725
AFG	country	2011	student	5291624	0.0023295
AFG	country	2012	student	5767543	0.0899382
AFG	country	2013	student	5986268	0.0379234
AFG	country	2014	student	6217756	0.0386698
AFG	country	2015	student	6199329	-0.0029636
AFG	country	2016	student	6265011	0.0105950
AFG	country	2017	student	6350404	0.0136301
AFG	country	2018	student	6544906	0.0306283
AFG	country	2019	student	6777785	0.0355817
AFG	country	2020	student	NA	NA
AFG	country	2021	student	NA	NA

kable(ppts_easia_weuro_long %>%
        filter(location_code == 'KOR_Busan' & variable == 'school'))

location_code	location_level	year	variable	value	pchg_yr1
KOR_Busan	province	1965	school	79	NA
KOR_Busan	province	1966	school	86	0.0886076
KOR_Busan	province	1967	school	93	0.0813953
KOR_Busan	province	1968	school	96	0.0322581
KOR_Busan	province	1969	school	98	0.0208333
KOR_Busan	province	1970	school	99	0.0102041
KOR_Busan	province	1971	school	100	0.0101010
KOR_Busan	province	1972	school	103	0.0300000
KOR_Busan	province	1973	school	106	0.0291262
KOR_Busan	province	1974	school	107	0.0094340
KOR_Busan	province	1975	school	112	0.0467290
KOR_Busan	province	1976	school	114	0.0178571
KOR_Busan	province	1977	school	114	0.0000000
KOR_Busan	province	1978	school	128	0.1228070
KOR_Busan	province	1979	school	131	0.0234375
KOR_Busan	province	1980	school	137	0.0458015
KOR_Busan	province	1981	school	149	0.0875912
KOR_Busan	province	1982	school	158	0.0604027
KOR_Busan	province	1983	school	169	0.0696203
KOR_Busan	province	1984	school	184	0.0887574
KOR_Busan	province	1985	school	193	0.0489130
KOR_Busan	province	1986	school	196	0.0155440
KOR_Busan	province	1987	school	200	0.0204082
KOR_Busan	province	1988	school	202	0.0100000
KOR_Busan	province	1989	school	217	0.0742574
KOR_Busan	province	1990	school	221	0.0184332
KOR_Busan	province	1991	school	222	0.0045249
KOR_Busan	province	1992	school	227	0.0225225
KOR_Busan	province	1993	school	230	0.0132159
KOR_Busan	province	1994	school	228	-0.0086957
KOR_Busan	province	1995	school	245	0.0745614
KOR_Busan	province	1996	school	250	0.0204082
KOR_Busan	province	1997	school	257	0.0280000
KOR_Busan	province	1998	school	259	0.0077821
KOR_Busan	province	1999	school	265	0.0231660
KOR_Busan	province	2000	school	267	0.0075472
KOR_Busan	province	2001	school	269	0.0074906
KOR_Busan	province	2002	school	273	0.0148699
KOR_Busan	province	2003	school	279	0.0219780
KOR_Busan	province	2004	school	283	0.0143369
KOR_Busan	province	2005	school	285	0.0070671
KOR_Busan	province	2006	school	292	0.0245614
KOR_Busan	province	2007	school	293	0.0034247
KOR_Busan	province	2008	school	293	0.0000000
KOR_Busan	province	2009	school	297	0.0136519
KOR_Busan	province	2010	school	298	0.0033670
KOR_Busan	province	2011	school	297	-0.0033557
KOR_Busan	province	2012	school	299	0.0067340
KOR_Busan	province	2013	school	302	0.0100334
KOR_Busan	province	2014	school	305	0.0099338
KOR_Busan	province	2015	school	306	0.0032787
KOR_Busan	province	2016	school	308	0.0065359
KOR_Busan	province	2017	school	308	0.0000000
KOR_Busan	province	2018	school	305	-0.0097403
KOR_Busan	province	2019	school	304	-0.0032787
KOR_Busan	province	2020	school	304	0.0000000
KOR_Busan	province	2021	school	304	0.0000000

Interpolating

We drop all NA values, only keeping the rows where we have observed levels. We then take the difference in levels between consecutive rows and divide by prior level to get percentage changes. This is potential percentage changes across multiple years if there were gaps with NA values.

We assume constant growth rate in the in-between years, and compute percentage annualized percentage changes. Given these, our annual percentage change formula is shown as below:

$$\text{annualPercentChange}_{\left(\text{from } t' \text{ to } t'+1\right)} = \left(\left( \frac{\text{schoolTeacherOrStudent}_{t+\tau}}{\text{schoolTeacherOrStudent}_{t}} \right)^{\frac{1}{\tau}} - 1\right),$$ where, the percentage change is for all $t \le t' \le t+\tau-1$.

The annual percentage change is exact where we know the level of schools, teachers, or students in the the current year and the year immediately after. But it is based growth trend “linear” interpolation when we have years of missing data in between.

We use the interpolated annual percentage changes to fill in gaps in levels. We interpolate with this function: ff_ppts_interp_linear. We test out the contents of the function via scripts below, and then confirm that the script and the function produce the same result. We first proceed with the script, and then the function.

Construct script-based results:

# D.2 annual_interp1
# D.2.1 compute linearly interpolated annual change
# create new sorting var, to take difference across years even if not
# consecutive
ppts_easia_weuro_interp1 <- ppts_easia_weuro_long %>%
  drop_na(value) %>%
  # New sorting variable, +1 regardless of year gap
  mutate(year_with_gap_ctr = row_number()) %>%
  arrange(location_code, location_level, variable, year_with_gap_ctr) %>%
  group_by(location_code, location_level, variable) %>%
  # compute percentage change over span and years gap over span
  mutate(pchg_span_interp1 = (value - lag(value)) / lag(value),
         span_yr = (year - lag(year))) %>%
  # linear-interpolated annualized (with compounding) percentage change
  mutate(pchg_yr1_interp1 = abs(1 + pchg_span_interp1) ^ (1 / span_yr) - 1) %>%
  # adjust values given percentages interpolated
  select(location_code, location_level,
         year, variable, span_yr, value, pchg_yr1_interp1)

# D.2.2 expand to all interpolating years
# the pchg_yr1_interp1 variable is only shown at start year, 
# due to drop_na(value) earlier
# need to expand to all years
ppts_easia_weuro_interp1 <- ppts_easia_weuro_interp1 %>%
  ungroup() %>% mutate(span_yr_dup = span_yr) %>%
  drop_na(span_yr) %>%
  tidyr::uncount(span_yr) %>%
  group_by(location_code, location_level, variable, year) %>%
  mutate(year_adj = row_number() + year - span_yr_dup, 
         gap_ctr = row_number()) %>%
  ungroup() %>%
  select(location_code, location_level,
         year_adj, variable, value, pchg_yr1_interp1, gap_ctr, span_yr_dup) %>%
  rename(year = year_adj)

# D.2.3, Fill in missing values
ppts_easia_weuro_interp1 <- ppts_easia_weuro_interp1 %>%
  arrange(location_code, location_level, variable, year) %>%
  group_by(location_code, location_level, variable) %>%
  mutate(value_interp1 = case_when(
    # Conditions A and B are both gap_ctr == span_yr_dup
    # A. value is correct, originally not NA
    gap_ctr == 1 & span_yr_dup == 1 ~ value,
    # B. value is also correct, next available year with non NA
    gap_ctr == span_yr_dup & span_yr_dup > 1 ~ value,
    # C. in-between years, where values were NA
    gap_ctr < span_yr_dup ~ (value/((1+pchg_yr1_interp1)^(span_yr_dup-gap_ctr)))
    )) %>%
  ungroup() %>%
  select(-gap_ctr, -span_yr_dup, -value)

# D.3 merge together
ppts_easia_weuro_interp1_script <- ppts_easia_weuro_long %>%
  left_join(ppts_easia_weuro_interp1,
            by = (c('location_code' = 'location_code',
                    'location_level' = 'location_level',
                    'year' = 'year',
                    'variable' = 'variable'))) %>%
  # add in data from the first year of raw data availability
  mutate(value_interp1 = case_when(
    !is.na(value) & is.na(value_interp1) ~ value,
    TRUE ~ value_interp1
    ))

# Print interpolated results
str(ppts_easia_weuro_interp1)
#> tibble [52,118 × 6] (S3: tbl_df/tbl/data.frame)
#>  $ location_code   : Factor w/ 286 levels "ABW","AFE","AFG",..: 1 1 1 1 1 1 1 1 1 1 ...
#>  $ location_level  : Factor w/ 4 levels "country","multicountry",..: 1 1 1 1 1 1 1 1 1 1 ...
#>  $ year            : num [1:52118] 1987 1988 1989 1990 1991 ...
#>  $ variable        : chr [1:52118] "gdp" "gdp" "gdp" "gdp" ...
#>  $ pchg_yr1_interp1: num [1:52118] 0.176 0.2014 0.122 0.0209 0.0384 ...
#>  $ value_interp1   : num [1:52118] 20263 24343 27313 27884 28954 ...
kable(ppts_easia_weuro_interp1 %>% select(-location_level) %>%
        filter(location_code == 'AFG' & variable == 'student'))

location_code	year	variable	pchg_yr1_interp1	value_interp1
AFG	1971	student	0.0596429	572933.0
AFG	1972	student	0.0428951	597509.0
AFG	1973	student	0.0449617	624374.0
AFG	1974	student	0.0477839	654209.0
AFG	1975	student	0.0582887	692342.0
AFG	1976	student	0.0539112	729667.0
AFG	1977	student	0.0472928	764175.0
AFG	1978	student	0.0608264	810657.0
AFG	1979	student	0.0879845	881982.2
AFG	1980	student	0.0879845	959583.0
AFG	1981	student	0.0679024	1024741.0
AFG	1982	student	-0.6433655	365458.0
AFG	1983	student	0.1063829	404336.5
AFG	1984	student	0.1063829	447351.0
AFG	1985	student	0.0710829	479150.0
AFG	1986	student	0.0702473	512809.0
AFG	1987	student	0.1272495	578063.7
AFG	1988	student	0.1272495	651622.0
AFG	1989	student	-0.0452517	622135.0
AFG	1990	student	0.0006076	622513.0
AFG	1991	student	0.0086344	627888.0
AFG	1992	student	0.1192242	702747.5
AFG	1993	student	0.1192242	786532.0
AFG	1994	student	0.4766646	1161444.0
AFG	1995	student	0.1297979	1312197.0
AFG	1996	student	-0.0726913	1216811.7
AFG	1997	student	-0.0726913	1128360.0
AFG	1998	student	-0.0726913	1046338.0
AFG	1999	student	-0.1631719	875605.0
AFG	2000	student	-0.1441803	749360.0
AFG	2001	student	0.0323783	773623.0
AFG	2002	student	2.4482287	2667629.0
AFG	2003	student	0.4173691	3781015.0
AFG	2004	student	0.1716806	4430142.0
AFG	2005	student	-0.0251285	4318819.0
AFG	2006	student	0.0811081	4669110.0
AFG	2007	student	0.0104874	4718077.0
AFG	2008	student	0.0544203	4974836.0
AFG	2009	student	-0.0058703	4945632.0
AFG	2010	student	0.0674725	5279326.0
AFG	2011	student	0.0023295	5291624.0
AFG	2012	student	0.0899382	5767543.0
AFG	2013	student	0.0379234	5986268.0
AFG	2014	student	0.0386698	6217756.0
AFG	2015	student	-0.0029636	6199329.0
AFG	2016	student	0.0105950	6265011.0
AFG	2017	student	0.0136301	6350404.0
AFG	2018	student	0.0306283	6544906.0
AFG	2019	student	0.0355817	6777785.0

Functionalizing the script, we have the function output below:

# Function parameters
df_data <- ppts_easia_weuro_long
ar_svr_group <- c("location_code", "location_level", "variable")
svr_data <- c("value")
svr_date <- c("year")
svr_interp <- c("value_interp1_func")
# Run function
ppts_easia_weuro_interp1_func <- PrjCompPPTS::ff_ppts_interp_linear(
  df_data,
  ar_svr_group = ar_svr_group,
  svr_data = svr_data,
  svr_date = svr_date,
  svr_interp = svr_interp,
  verbose = FALSE
) %>% rename(value_interp1 = value_interp1_func)
# Test if identical
bl_func_script_consistency <- identical(
  ppts_easia_weuro_interp1_func, ppts_easia_weuro_interp1_script)
print(glue::glue("bl_func_script_consistency = {bl_func_script_consistency}"))
#> bl_func_script_consistency = TRUE

Year selection:

# Only consider years in this range
it_min_year <- 1920
it_max_year <- 2020
ppts_easia_weuro_long <- ppts_easia_weuro_interp1_func
ppts_easia_weuro_long <- ppts_easia_weuro_long %>%
    filter(year <= it_max_year & year >= it_min_year)

Extrapolating

We will extrapolate for several years before the start and after the end of the data timeframes. Extrapolation will not exceed going 5 years forward and going 5 years backwards. And extrapolation will only happen within years in which there is at least one variable, among variables for the country, has non-NA values. In cases where a country’s data is available only starting after 1980, we allow for extrapolation back to up to 1980 for 5 years.

Specifically, because we generally have population data from 1960 to 2020 for all countries, we will not be extrapolating prior to 1960 or after 2020. But data for Korea starts in 1965, so we will not extrapolate to any years before 1965, but if one of the Korean variables has data starting from 1970, we will extrapolate between 1965 and 1970. But if the other variable only has data starting in 1980, we will extrapolate at most five years, back to 1975. For Germany, unification happened in 1992. We do not have data in 1990, preventing us from computing change from 1990 to 2000. We extrapolate from 1992 back 5 years to 1987, generating a value for 1990.

Extrapolation is meant to help with situations where we have data up to 2019, but for consistency of comparison, would be useful to extend the data to 2020 by extrapolating 1 year forward.

Compute start and end year for each variable and each country. For the interpolated value column, NaN values are before and after the start and end of available data. First go forward, value = lag(value) x lag(1+change), fill in if value is NaN and current year is less than 5 above terimal data availability time. Then do the same going backwards, but there, value = lead(value, n=1) / lead(1 + change, n=2)

First, we generate end and start years and interpolating percentages, etc.

# Generate start and end times
ppts_easia_weuro_extrapolate <- ppts_easia_weuro_long %>%
  drop_na(value) %>%
  arrange(location_code, location_level, variable, year) %>%
  group_by(location_code, location_level, variable) %>%
  mutate(year_start = first(year, na_rm=TRUE), year_end = last(year, na_rm=TRUE),
         value_start = first(value, na_rm=TRUE), value_end = last(value, na_rm=TRUE),
         pchg_yr1_interp1_start = first(pchg_yr1_interp1, na_rm=TRUE),
         pchg_yr1_interp1_end = last(pchg_yr1_interp1, na_rm=TRUE)) %>% 
  slice(1) %>%
  select(location_code, location_level, variable, 
         contains("_start"), contains("_end"))

Second extrapolate.

# How many years to extrapolate back
it_year_extrplt_backmost <- 5
it_year_extrplt_fordmost <- 5

# Merge with full skeleton for expansion
ppts_easia_weuro_extrapolate <- ppts_easia_weuro_long %>%
  select(location_code, location_level, variable, year, value) %>%
  left_join(ppts_easia_weuro_extrapolate,
            by = (c('location_code' = 'location_code',
                    'location_level' = 'location_level',
                    'variable' = 'variable'))) 

# Extrapolate forward
ppts_easia_weuro_extrapolate <- ppts_easia_weuro_extrapolate %>%
  mutate(value_extrapolate =
    case_when(
        is.na(value) & (year_end + it_year_extrplt_fordmost >= year) & (year > year_end) ~
          (value_end*((1+pchg_yr1_interp1_end)^(year-year_end))),
        is.na(value) & (year_start - it_year_extrplt_backmost <= year) & (year < year_start) ~
          (value_start/((1+pchg_yr1_interp1_start)^(year_start - year))),
        TRUE ~ NA
      )) %>%
  mutate(pchg_yr1_interp1_extrapolate =
    case_when(
        (year_end + it_year_extrplt_fordmost >= year) & (year >= year_end) ~
          pchg_yr1_interp1_end,
        (year_start - it_year_extrplt_backmost <= year) & (year <= year_start) ~
          pchg_yr1_interp1_start,
        TRUE ~ NA
      ))    

# Get extrapolated values
ppts_easia_weuro_extrapolate <- ppts_easia_weuro_extrapolate %>%
  drop_na(pchg_yr1_interp1_extrapolate) %>% 
  select(location_code, location_level, variable, year, 
         value_extrapolate, pchg_yr1_interp1_extrapolate) %>%
  rename(value_interp1 = value_extrapolate, 
         pchg_yr1_interp1 = pchg_yr1_interp1_extrapolate)

# Print
kable(ppts_easia_weuro_extrapolate %>% select(-location_level) %>%
        filter(location_code == 'AFG' & variable == 'student'))
#> Adding missing grouping variables: `location_level`

location_level	location_code	variable	year	value_interp1	pchg_yr1_interp1
country	AFG	student	1965	404712.6	0.0596429
country	AFG	student	1966	428850.8	0.0596429
country	AFG	student	1967	454428.7	0.0596429
country	AFG	student	1968	481532.2	0.0596429
country	AFG	student	1969	510252.1	0.0596429
country	AFG	student	1970	NA	0.0596429
country	AFG	student	2019	NA	0.0355817
country	AFG	student	2020	7018950.2	0.0355817

# kable(ppts_easia_weuro_extrapolate %>% select(-location_level) %>%
#         filter(location_code == 'KOR_Busan' & variable == 'school'))

Merge raw and interpolated results together with extrapolated results

Merge extrpolated results back to main dataframe.

# Merge
ppts_easia_weuro_long <- ppts_easia_weuro_long %>%
  left_join(ppts_easia_weuro_extrapolate,
            by = (c('location_code' = 'location_code',
                    'location_level' = 'location_level',
                    'year' = 'year',
                    'variable' = 'variable'))) 

# Combine columns
ppts_easia_weuro_long <- ppts_easia_weuro_long %>%            
  mutate(value_interp1 = coalesce(value_interp1.x, value_interp1.y)) %>%
  select(-value_interp1.x, -value_interp1.y) %>%
  mutate(pchg_yr1_interp1 = coalesce(pchg_yr1_interp1.x, pchg_yr1_interp1.y)) %>%
  select(-pchg_yr1_interp1.x, -pchg_yr1_interp1.y) %>%
  drop_na(value_interp1)

Interpolation and Extrapolation Results Overview

The results in the value_interp1 column below contains below interpolation and extrapolation results.

Afghanistan Interpolate and Extrapolate Results

We illustrate the interpolation results by showing outputs from Afghanistan for student counts. Here, we are extrapolating several years back to 1965, and forward 1 year to 2020. And we fill in several years of missing data.

By extrapolating data in 2020, we are able to generate percentage change between 2000 and 2020, 2010 and 2020, etc. We are also able to see levels in 2020. These percentages and levels were not available without interpolation/extrapolation. We can see that our data for Afghanistan is much more complete after interpolating/extrapolating.

# str(ppts_easia_weuro_long)
# print
kable(ppts_easia_weuro_long %>% 
        filter(location_code == 'AFG' & variable == 'student') %>%
        ungroup() %>%
        select(location_code, year, variable, 
          value, value_interp1, pchg_yr1, pchg_yr1_interp1), 
      caption="Raw and interpolated Afghanistan students results")

Raw and interpolated Afghanistan students results

location_code	year	variable	value	value_interp1	pchg_yr1	pchg_yr1_interp1
AFG	1965	student	NA	404712.6	NA	0.0596429
AFG	1966	student	NA	428850.8	NA	0.0596429
AFG	1967	student	NA	454428.7	NA	0.0596429
AFG	1968	student	NA	481532.2	NA	0.0596429
AFG	1969	student	NA	510252.1	NA	0.0596429
AFG	1970	student	540685	540685.0	NA	0.0596429
AFG	1971	student	572933	572933.0	0.0596429	0.0596429
AFG	1972	student	597509	597509.0	0.0428951	0.0428951
AFG	1973	student	624374	624374.0	0.0449617	0.0449617
AFG	1974	student	654209	654209.0	0.0477839	0.0477839
AFG	1975	student	692342	692342.0	0.0582887	0.0582887
AFG	1976	student	729667	729667.0	0.0539112	0.0539112
AFG	1977	student	764175	764175.0	0.0472928	0.0472928
AFG	1978	student	810657	810657.0	0.0608264	0.0608264
AFG	1979	student	NA	881982.2	NA	0.0879845
AFG	1980	student	959583	959583.0	NA	0.0879845
AFG	1981	student	1024741	1024741.0	0.0679024	0.0679024
AFG	1982	student	365458	365458.0	-0.6433655	-0.6433655
AFG	1983	student	NA	404336.5	NA	0.1063829
AFG	1984	student	447351	447351.0	NA	0.1063829
AFG	1985	student	479150	479150.0	0.0710829	0.0710829
AFG	1986	student	512809	512809.0	0.0702473	0.0702473
AFG	1987	student	NA	578063.7	NA	0.1272495
AFG	1988	student	651622	651622.0	NA	0.1272495
AFG	1989	student	622135	622135.0	-0.0452517	-0.0452517
AFG	1990	student	622513	622513.0	0.0006076	0.0006076
AFG	1991	student	627888	627888.0	0.0086344	0.0086344
AFG	1992	student	NA	702747.5	NA	0.1192242
AFG	1993	student	786532	786532.0	NA	0.1192242
AFG	1994	student	1161444	1161444.0	0.4766646	0.4766646
AFG	1995	student	1312197	1312197.0	0.1297979	0.1297979
AFG	1996	student	NA	1216811.7	NA	-0.0726913
AFG	1997	student	NA	1128360.0	NA	-0.0726913
AFG	1998	student	1046338	1046338.0	NA	-0.0726913
AFG	1999	student	875605	875605.0	-0.1631719	-0.1631719
AFG	2000	student	749360	749360.0	-0.1441803	-0.1441803
AFG	2001	student	773623	773623.0	0.0323783	0.0323783
AFG	2002	student	2667629	2667629.0	2.4482287	2.4482287
AFG	2003	student	3781015	3781015.0	0.4173691	0.4173691
AFG	2004	student	4430142	4430142.0	0.1716806	0.1716806
AFG	2005	student	4318819	4318819.0	-0.0251285	-0.0251285
AFG	2006	student	4669110	4669110.0	0.0811081	0.0811081
AFG	2007	student	4718077	4718077.0	0.0104874	0.0104874
AFG	2008	student	4974836	4974836.0	0.0544203	0.0544203
AFG	2009	student	4945632	4945632.0	-0.0058703	-0.0058703
AFG	2010	student	5279326	5279326.0	0.0674725	0.0674725
AFG	2011	student	5291624	5291624.0	0.0023295	0.0023295
AFG	2012	student	5767543	5767543.0	0.0899382	0.0899382
AFG	2013	student	5986268	5986268.0	0.0379234	0.0379234
AFG	2014	student	6217756	6217756.0	0.0386698	0.0386698
AFG	2015	student	6199329	6199329.0	-0.0029636	-0.0029636
AFG	2016	student	6265011	6265011.0	0.0105950	0.0105950
AFG	2017	student	6350404	6350404.0	0.0136301	0.0136301
AFG	2018	student	6544906	6544906.0	0.0306283	0.0306283
AFG	2019	student	6777785	6777785.0	0.0355817	0.0355817
AFG	2020	student	NA	7018950.2	NA	0.0355817

Austria Interpolate and Extrapolate Results

The Austrian data starts in 1923, and school count is available in that year. Hence, we do not extrapolate prior to that. If the data starts in 1923 (meaning any measures of any information is available in 1923), but the school count data starts in 1925, we would extrapolate back 2 years to 1923. Austrian data for initial decades is spotty, hence we interpolate to fill in the gaps.

# str(ppts_easia_weuro_long)
# print
kable(ppts_easia_weuro_long %>% 
        filter(location_code == 'AUT' & variable == 'school') %>%
        ungroup() %>%
        select(location_code, year, variable, 
          value, value_interp1, pchg_yr1, pchg_yr1_interp1), 
      caption="Raw and interpolated Austria schools results")

Raw and interpolated Austria schools results

location_code	year	variable	value	value_interp1	pchg_yr1	pchg_yr1_interp1
AUT	1923	school	4655	4655.000	NA	0.0045011
AUT	1924	school	NA	4675.953	NA	0.0045011
AUT	1925	school	4697	4697.000	NA	0.0045011
AUT	1926	school	NA	4697.400	NA	0.0000851
AUT	1927	school	NA	4697.800	NA	0.0000851
AUT	1928	school	NA	4698.200	NA	0.0000851
AUT	1929	school	NA	4698.600	NA	0.0000851
AUT	1930	school	4699	4699.000	NA	0.0000851
AUT	1931	school	NA	4693.989	NA	-0.0010663
AUT	1932	school	NA	4688.984	NA	-0.0010663
AUT	1933	school	NA	4683.984	NA	-0.0010663
AUT	1934	school	NA	4678.989	NA	-0.0010663
AUT	1935	school	4674	4674.000	NA	-0.0010663
AUT	1936	school	NA	4634.460	NA	-0.0084595
AUT	1937	school	NA	4595.255	NA	-0.0084595
AUT	1938	school	NA	4556.381	NA	-0.0084595
AUT	1939	school	NA	4517.836	NA	-0.0084595
AUT	1940	school	NA	4479.618	NA	-0.0084595
AUT	1941	school	NA	4441.722	NA	-0.0084595
AUT	1942	school	NA	4404.147	NA	-0.0084595
AUT	1943	school	NA	4366.891	NA	-0.0084595
AUT	1944	school	NA	4329.949	NA	-0.0084595
AUT	1945	school	NA	4293.319	NA	-0.0084595
AUT	1946	school	4257	4257.000	NA	-0.0084595
AUT	1947	school	NA	4292.308	NA	0.0082942
AUT	1948	school	NA	4327.909	NA	0.0082942
AUT	1949	school	NA	4363.806	NA	0.0082942
AUT	1950	school	4400	4400.000	NA	0.0082942
AUT	1951	school	4417	4417.000	0.0038636	0.0038636
AUT	1952	school	4417	4417.000	0.0000000	0.0000000
AUT	1953	school	4426	4426.000	0.0020376	0.0020376
AUT	1954	school	4426	4426.000	0.0000000	0.0000000
AUT	1955	school	4427	4427.000	0.0002259	0.0002259
AUT	1956	school	4426	4426.000	-0.0002259	-0.0002259
AUT	1957	school	4418	4418.000	-0.0018075	-0.0018075
AUT	1958	school	4403	4403.000	-0.0033952	-0.0033952
AUT	1959	school	4402	4402.000	-0.0002271	-0.0002271
AUT	1960	school	4393	4393.000	-0.0020445	-0.0020445
AUT	1961	school	4387	4387.000	-0.0013658	-0.0013658
AUT	1962	school	4374	4374.000	-0.0029633	-0.0029633
AUT	1963	school	4374	4374.000	0.0000000	0.0000000
AUT	1964	school	4375	4375.000	0.0002286	0.0002286
AUT	1965	school	4295	4295.000	-0.0182857	-0.0182857
AUT	1966	school	4175	4175.000	-0.0279395	-0.0279395
AUT	1967	school	4094	4094.000	-0.0194012	-0.0194012
AUT	1968	school	4059	4059.000	-0.0085491	-0.0085491
AUT	1969	school	4018	4018.000	-0.0101010	-0.0101010
AUT	1970	school	3973	3973.000	-0.0111996	-0.0111996
AUT	1971	school	3839	3839.000	-0.0337277	-0.0337277
AUT	1972	school	3783	3783.000	-0.0145871	-0.0145871
AUT	1973	school	3711	3711.000	-0.0190325	-0.0190325
AUT	1974	school	3644	3644.000	-0.0180544	-0.0180544
AUT	1975	school	3590	3590.000	-0.0148189	-0.0148189
AUT	1976	school	3548	3548.000	-0.0116992	-0.0116992
AUT	1977	school	3508	3508.000	-0.0112740	-0.0112740
AUT	1978	school	3494	3494.000	-0.0039909	-0.0039909
AUT	1979	school	3466	3466.000	-0.0080137	-0.0080137
AUT	1980	school	3450	3450.000	-0.0046163	-0.0046163
AUT	1981	school	3451	3451.000	0.0002899	0.0002899
AUT	1982	school	3434	3434.000	-0.0049261	-0.0049261
AUT	1983	school	3421	3421.000	-0.0037857	-0.0037857
AUT	1984	school	3414	3414.000	-0.0020462	-0.0020462
AUT	1985	school	3411	3411.000	-0.0008787	-0.0008787
AUT	1986	school	3395	3395.000	-0.0046907	-0.0046907
AUT	1987	school	3394	3394.000	-0.0002946	-0.0002946
AUT	1988	school	3385	3385.000	-0.0026517	-0.0026517
AUT	1989	school	3383	3383.000	-0.0005908	-0.0005908
AUT	1990	school	3386	3386.000	0.0008868	0.0008868
AUT	1991	school	3384	3384.000	-0.0005907	-0.0005907
AUT	1992	school	3381	3381.000	-0.0008865	-0.0008865
AUT	1993	school	3382	3382.000	0.0002958	0.0002958
AUT	1994	school	3378	3378.000	-0.0011827	-0.0011827
AUT	1995	school	3383	3383.000	0.0014802	0.0014802
AUT	1996	school	3367	3367.000	-0.0047295	-0.0047295
AUT	1997	school	3362	3362.000	-0.0014850	-0.0014850
AUT	1998	school	3366	3366.000	0.0011898	0.0011898
AUT	1999	school	3364	3364.000	-0.0005942	-0.0005942
AUT	2000	school	3360	3360.000	-0.0011891	-0.0011891
AUT	2001	school	3309	3309.000	-0.0151786	-0.0151786
AUT	2002	school	3299	3299.000	-0.0030221	-0.0030221
AUT	2003	school	3336	3336.000	0.0112155	0.0112155
AUT	2004	school	3324	3324.000	-0.0035971	-0.0035971
AUT	2005	school	3296	3296.000	-0.0084236	-0.0084236
AUT	2006	school	3248	3248.000	-0.0145631	-0.0145631
AUT	2007	school	3225	3225.000	-0.0070813	-0.0070813
AUT	2008	school	3207	3207.000	-0.0055814	-0.0055814
AUT	2009	school	3197	3197.000	-0.0031182	-0.0031182
AUT	2010	school	3171	3171.000	-0.0081326	-0.0081326
AUT	2011	school	3135	3135.000	-0.0113529	-0.0113529
AUT	2012	school	3095	3095.000	-0.0127592	-0.0127592
AUT	2013	school	3066	3066.000	-0.0093700	-0.0093700
AUT	2014	school	3051	3051.000	-0.0048924	-0.0048924
AUT	2015	school	3039	3039.000	-0.0039331	-0.0039331
AUT	2016	school	3040	3040.000	0.0003291	0.0003291
AUT	2017	school	3033	3033.000	-0.0023026	-0.0023026
AUT	2018	school	3026	3026.000	-0.0023079	-0.0023079
AUT	2019	school	3014	3014.000	-0.0039656	-0.0039656
AUT	2020	school	3014	3014.000	0.0000000	0.0000000

Switzerland Example

We have some historical data on Switzerland, and also some recent data.

# str(ppts_easia_weuro_long)
# print
kable(ppts_easia_weuro_long %>% 
        filter(location_code == 'CHE' & variable == 'teacher') %>%
        ungroup() %>%
        select(location_code, year, variable, 
          value, value_interp1, pchg_yr1, pchg_yr1_interp1), 
      caption="Raw and interpolated Switzerland schools results")

Raw and interpolated Switzerland schools results

location_code	year	variable	value	value_interp1	pchg_yr1	pchg_yr1_interp1
CHE	1920	teacher	NA	13497.39	NA	-0.0023365
CHE	1921	teacher	NA	13465.85	NA	-0.0023365
CHE	1922	teacher	NA	13434.39	NA	-0.0023365
CHE	1923	teacher	13403	13403.00	NA	-0.0023365
CHE	1924	teacher	NA	13345.89	NA	-0.0042613
CHE	1925	teacher	NA	13289.02	NA	-0.0042613
CHE	1926	teacher	NA	13232.39	NA	-0.0042613
CHE	1927	teacher	13176	13176.00	NA	-0.0042613
CHE	1928	teacher	NA	13207.88	NA	0.0024199
CHE	1929	teacher	NA	13239.85	NA	0.0024199
CHE	1930	teacher	NA	13271.88	NA	0.0024199
CHE	1931	teacher	13304	13304.00	NA	0.0024199
CHE	1932	teacher	NA	13371.73	NA	0.0050910
CHE	1933	teacher	NA	13439.81	NA	0.0050910
CHE	1934	teacher	NA	13508.23	NA	0.0050910
CHE	1935	teacher	13577	13577.00	NA	0.0050910
CHE	1936	teacher	NA	13572.50	NA	-0.0003316
CHE	1937	teacher	NA	13568.00	NA	-0.0003316
CHE	1938	teacher	NA	13563.50	NA	-0.0003316
CHE	1939	teacher	13559	13559.00	NA	-0.0003316
CHE	1940	teacher	NA	13533.93	NA	-0.0018489
CHE	1941	teacher	NA	13508.91	NA	-0.0018489
CHE	1942	teacher	NA	13483.93	NA	-0.0018489
CHE	1943	teacher	13459	13459.00	NA	-0.0018489
CHE	1944	teacher	NA	13575.72	NA	0.0086725
CHE	1945	teacher	NA	13693.46	NA	0.0086725
CHE	1946	teacher	NA	13812.21	NA	0.0086725
CHE	1947	teacher	13932	13932.00	NA	0.0086725
CHE	1948	teacher	NA	14066.05	NA	0.0096219
CHE	1949	teacher	NA	14201.40	NA	0.0096219
CHE	1950	teacher	NA	14338.04	NA	0.0096219
CHE	1951	teacher	14476	14476.00	NA	0.0096219
CHE	1952	teacher	NA	14847.08	NA	0.0256342
CHE	1953	teacher	NA	15227.67	NA	0.0256342
CHE	1954	teacher	NA	15618.02	NA	0.0256342
CHE	1955	teacher	NA	16018.38	NA	0.0256342
CHE	1956	teacher	16429	16429.00	NA	0.0256342
CHE	1957	teacher	NA	16678.32	NA	0.0151754
CHE	1958	teacher	NA	16931.42	NA	0.0151754
CHE	1959	teacher	NA	17188.36	NA	0.0151754
CHE	1960	teacher	NA	17449.20	NA	0.0151754
CHE	1961	teacher	17714	17714.00	NA	0.0151754
CHE	1962	teacher	NA	18046.84	NA	0.0187896
CHE	1963	teacher	NA	18385.93	NA	0.0187896
CHE	1964	teacher	NA	18731.40	NA	0.0187896
CHE	1965	teacher	NA	19083.36	NA	0.0187896
CHE	1966	teacher	NA	19441.92	NA	0.0187896
CHE	1967	teacher	NA	19807.23	NA	0.0187896
CHE	1968	teacher	NA	20179.40	NA	0.0187896
CHE	1969	teacher	NA	20558.57	NA	0.0187896
CHE	1970	teacher	NA	20944.85	NA	0.0187896
CHE	1971	teacher	NA	21338.40	NA	0.0187896
CHE	1972	teacher	NA	21739.34	NA	0.0187896
CHE	1973	teacher	NA	22147.82	NA	0.0187896
CHE	1974	teacher	NA	22563.97	NA	0.0187896
CHE	1975	teacher	NA	22987.94	NA	0.0187896
CHE	1976	teacher	NA	23419.87	NA	0.0187896
CHE	1977	teacher	NA	23859.92	NA	0.0187896
CHE	1978	teacher	NA	24308.24	NA	0.0187896
CHE	1979	teacher	NA	24764.98	NA	0.0187896
CHE	1980	teacher	NA	25230.31	NA	0.0187896
CHE	1981	teacher	NA	25704.38	NA	0.0187896
CHE	1982	teacher	NA	26187.35	NA	0.0187896
CHE	1983	teacher	NA	26679.41	NA	0.0187896
CHE	1984	teacher	NA	27180.70	NA	0.0187896
CHE	1985	teacher	NA	27691.42	NA	0.0187896
CHE	1986	teacher	NA	28211.73	NA	0.0187896
CHE	1987	teacher	NA	28741.82	NA	0.0187896
CHE	1988	teacher	NA	29281.87	NA	0.0187896
CHE	1989	teacher	NA	29832.06	NA	0.0187896
CHE	1990	teacher	NA	30392.60	NA	0.0187896
CHE	1991	teacher	NA	30963.66	NA	0.0187896
CHE	1992	teacher	NA	31545.46	NA	0.0187896
CHE	1993	teacher	NA	32138.19	NA	0.0187896
CHE	1994	teacher	NA	32742.05	NA	0.0187896
CHE	1995	teacher	NA	33357.26	NA	0.0187896
CHE	1996	teacher	NA	33984.04	NA	0.0187896
CHE	1997	teacher	NA	34622.58	NA	0.0187896
CHE	1998	teacher	NA	35273.13	NA	0.0187896
CHE	1999	teacher	NA	35935.90	NA	0.0187896
CHE	2000	teacher	NA	36611.12	NA	0.0187896
CHE	2001	teacher	NA	37299.03	NA	0.0187896
CHE	2002	teacher	NA	37999.87	NA	0.0187896
CHE	2003	teacher	NA	38713.87	NA	0.0187896
CHE	2004	teacher	NA	39441.29	NA	0.0187896
CHE	2005	teacher	NA	40182.38	NA	0.0187896
CHE	2006	teacher	NA	40937.39	NA	0.0187896
CHE	2007	teacher	NA	41706.59	NA	0.0187896
CHE	2008	teacher	NA	42490.25	NA	0.0187896
CHE	2009	teacher	NA	43288.62	NA	0.0187896
CHE	2010	teacher	44102	44102.00	NA	0.0187896
CHE	2011	teacher	45666	45666.00	0.0354632	0.0354632
CHE	2012	teacher	46104	46104.00	0.0095914	0.0095914
CHE	2013	teacher	48345	48345.00	0.0486075	0.0486075
CHE	2014	teacher	47223	47223.00	-0.0232082	-0.0232082
CHE	2015	teacher	48595	48595.00	0.0290536	0.0290536
CHE	2016	teacher	50672	50672.00	0.0427410	0.0427410
CHE	2017	teacher	50879	50879.00	0.0040851	0.0040851
CHE	2018	teacher	51844	51844.00	0.0189666	0.0189666
CHE	2019	teacher	52738	52738.00	0.0172440	0.0172440
CHE	2020	teacher	54777	54777.00	0.0386628	0.0386628

Germany Example

Germany reunified in 1992, we extrapolate to obtain data predictions for 1990 to facilitate computing percentage changes.

# str(ppts_easia_weuro_long)
# print
kable(ppts_easia_weuro_long %>% 
        filter(location_code == 'DEU' & variable == 'school') %>%
        ungroup() %>%
        select(location_code, year, variable, 
          value, value_interp1, pchg_yr1, pchg_yr1_interp1), 
      caption="Raw and interpolated Germany schools results")

Raw and interpolated Germany schools results

location_code	year	variable	value	value_interp1	pchg_yr1	pchg_yr1_interp1
DEU	1987	school	NA	18091.76	NA	-0.0016721
DEU	1988	school	NA	18061.50	NA	-0.0016721
DEU	1989	school	NA	18031.30	NA	-0.0016721
DEU	1990	school	NA	18001.15	NA	-0.0016721
DEU	1991	school	NA	17971.05	NA	-0.0016721
DEU	1992	school	17941	17941.00	NA	-0.0016721
DEU	1993	school	17911	17911.00	-0.0016721	-0.0016721
DEU	1994	school	17895	17895.00	-0.0008933	-0.0008933
DEU	1995	school	17910	17910.00	0.0008382	0.0008382
DEU	1996	school	17892	17892.00	-0.0010050	-0.0010050
DEU	1997	school	17829	17829.00	-0.0035211	-0.0035211
DEU	1998	school	17662	17662.00	-0.0093668	-0.0093668
DEU	1999	school	17503	17503.00	-0.0090024	-0.0090024
DEU	2000	school	17275	17275.00	-0.0130263	-0.0130263
DEU	2001	school	17175	17175.00	-0.0057887	-0.0057887
DEU	2002	school	17075	17075.00	-0.0058224	-0.0058224
DEU	2003	school	16992	16992.00	-0.0048609	-0.0048609
DEU	2004	school	16932	16932.00	-0.0035311	-0.0035311
DEU	2005	school	16814	16814.00	-0.0069691	-0.0069691
DEU	2006	school	16743	16743.00	-0.0042227	-0.0042227
DEU	2007	school	16649	16649.00	-0.0056143	-0.0056143
DEU	2008	school	16500	16500.00	-0.0089495	-0.0089495
DEU	2009	school	16431	16431.00	-0.0041818	-0.0041818
DEU	2010	school	16290	16290.00	-0.0085813	-0.0085813
DEU	2011	school	16103	16103.00	-0.0114794	-0.0114794
DEU	2012	school	15971	15971.00	-0.0081972	-0.0081972
DEU	2013	school	15749	15749.00	-0.0139002	-0.0139002
DEU	2014	school	15578	15578.00	-0.0108578	-0.0108578
DEU	2015	school	15483	15483.00	-0.0060983	-0.0060983
DEU	2016	school	15456	15456.00	-0.0017438	-0.0017438
DEU	2017	school	15409	15409.00	-0.0030409	-0.0030409
DEU	2018	school	15398	15398.00	-0.0007139	-0.0007139
DEU	2019	school	15431	15431.00	0.0021431	0.0021431
DEU	2020	school	15447	15447.00	0.0010369	0.0010369

Korean Busan Interpolate and Extrapolate Results

We do not do any extrapolation or interpolation in Busan Korea, because the data is available for school counts between 1965 and 2020, the min and max years of the Korean data, and data is available every single year.

# str(ppts_easia_weuro_long)
# print
kable(ppts_easia_weuro_long %>% 
        filter(location_code == 'KOR_Busan' & variable == 'school') %>%
        ungroup() %>%
        select(location_code, year, variable, 
          value, value_interp1, pchg_yr1, pchg_yr1_interp1), 
      caption="Raw and interpolated Korean Busan province schools results")

Raw and interpolated Korean Busan province schools results

location_code	year	variable	value	value_interp1	pchg_yr1	pchg_yr1_interp1
KOR_Busan	1965	school	79	79	NA	0.0886076
KOR_Busan	1966	school	86	86	0.0886076	0.0886076
KOR_Busan	1967	school	93	93	0.0813953	0.0813953
KOR_Busan	1968	school	96	96	0.0322581	0.0322581
KOR_Busan	1969	school	98	98	0.0208333	0.0208333
KOR_Busan	1970	school	99	99	0.0102041	0.0102041
KOR_Busan	1971	school	100	100	0.0101010	0.0101010
KOR_Busan	1972	school	103	103	0.0300000	0.0300000
KOR_Busan	1973	school	106	106	0.0291262	0.0291262
KOR_Busan	1974	school	107	107	0.0094340	0.0094340
KOR_Busan	1975	school	112	112	0.0467290	0.0467290
KOR_Busan	1976	school	114	114	0.0178571	0.0178571
KOR_Busan	1977	school	114	114	0.0000000	0.0000000
KOR_Busan	1978	school	128	128	0.1228070	0.1228070
KOR_Busan	1979	school	131	131	0.0234375	0.0234375
KOR_Busan	1980	school	137	137	0.0458015	0.0458015
KOR_Busan	1981	school	149	149	0.0875912	0.0875912
KOR_Busan	1982	school	158	158	0.0604027	0.0604027
KOR_Busan	1983	school	169	169	0.0696203	0.0696203
KOR_Busan	1984	school	184	184	0.0887574	0.0887574
KOR_Busan	1985	school	193	193	0.0489130	0.0489130
KOR_Busan	1986	school	196	196	0.0155440	0.0155440
KOR_Busan	1987	school	200	200	0.0204082	0.0204082
KOR_Busan	1988	school	202	202	0.0100000	0.0100000
KOR_Busan	1989	school	217	217	0.0742574	0.0742574
KOR_Busan	1990	school	221	221	0.0184332	0.0184332
KOR_Busan	1991	school	222	222	0.0045249	0.0045249
KOR_Busan	1992	school	227	227	0.0225225	0.0225225
KOR_Busan	1993	school	230	230	0.0132159	0.0132159
KOR_Busan	1994	school	228	228	-0.0086957	-0.0086957
KOR_Busan	1995	school	245	245	0.0745614	0.0745614
KOR_Busan	1996	school	250	250	0.0204082	0.0204082
KOR_Busan	1997	school	257	257	0.0280000	0.0280000
KOR_Busan	1998	school	259	259	0.0077821	0.0077821
KOR_Busan	1999	school	265	265	0.0231660	0.0231660
KOR_Busan	2000	school	267	267	0.0075472	0.0075472
KOR_Busan	2001	school	269	269	0.0074906	0.0074906
KOR_Busan	2002	school	273	273	0.0148699	0.0148699
KOR_Busan	2003	school	279	279	0.0219780	0.0219780
KOR_Busan	2004	school	283	283	0.0143369	0.0143369
KOR_Busan	2005	school	285	285	0.0070671	0.0070671
KOR_Busan	2006	school	292	292	0.0245614	0.0245614
KOR_Busan	2007	school	293	293	0.0034247	0.0034247
KOR_Busan	2008	school	293	293	0.0000000	0.0000000
KOR_Busan	2009	school	297	297	0.0136519	0.0136519
KOR_Busan	2010	school	298	298	0.0033670	0.0033670
KOR_Busan	2011	school	297	297	-0.0033557	-0.0033557
KOR_Busan	2012	school	299	299	0.0067340	0.0067340
KOR_Busan	2013	school	302	302	0.0100334	0.0100334
KOR_Busan	2014	school	305	305	0.0099338	0.0099338
KOR_Busan	2015	school	306	306	0.0032787	0.0032787
KOR_Busan	2016	school	308	308	0.0065359	0.0065359
KOR_Busan	2017	school	308	308	0.0000000	0.0000000
KOR_Busan	2018	school	305	305	-0.0097403	-0.0097403
KOR_Busan	2019	school	304	304	-0.0032787	-0.0032787
KOR_Busan	2020	school	304	304	0.0000000	0.0000000

Generate percentage changes every 5, 10, 15, 20 years

We now consider several different cuts, over the end-points of which we compute percentage changes.

# Every five years from 1940 until 2020
ar_it_cuts_1940t2020_i05 <- seq(1920, 2020, length.out = 21)
st_cuts_1940t2020_i05 <- "1920t2020i05"
# Every ten years from 1940 until 2020
ar_it_cuts_1940t2020_i10 <- seq(1920, 2020, length.out = 11)
st_cuts_1940t2020_i10 <- "1920t2020i10"
# Every 15 years from 1940 until 2020
ar_it_cuts_1940t2020_i15 <- seq(1925, 2015, length.out = 7)
st_cuts_1940t2020_i15 <- "1925t2015i15"
# Every 20 years from 1940 until 2020
ar_it_cuts_1940t2020_i20 <- seq(1920, 2020, length.out = 6)
st_cuts_1940t2020_i20 <- "1920t2020i20"

We put the cuts and the associated string names into two lists.

# List of cuts
ls_ar_it_cuts <- list(
  ar_it_cuts_1940t2020_i05,
  ar_it_cuts_1940t2020_i10,
  ar_it_cuts_1940t2020_i15,
  ar_it_cuts_1940t2020_i20
)
# Add names
names(ls_ar_it_cuts) <- c(
  st_cuts_1940t2020_i05,
  st_cuts_1940t2020_i10,
  st_cuts_1940t2020_i15,
  st_cuts_1940t2020_i20
)
# Display
for (st_cuts_name in names(ls_ar_it_cuts)) {
  print(glue::glue(
    "cutTypeName={st_cuts_name}:\n",
    "bins={ls_ar_it_cuts[st_cuts_name]}"))
}
#> cutTypeName=1920t2020i05:
#> bins=c(1920, 1925, 1930, 1935, 1940, 1945, 1950, 1955, 1960, 1965, 1970, 1975, 1980, 1985, 1990, 1995, 2000, 2005, 2010, 2015, 2020)
#> cutTypeName=1920t2020i10:
#> bins=c(1920, 1930, 1940, 1950, 1960, 1970, 1980, 1990, 2000, 2010, 2020)
#> cutTypeName=1925t2015i15:
#> bins=c(1925, 1940, 1955, 1970, 1985, 2000, 2015)
#> cutTypeName=1920t2020i20:
#> bins=c(1920, 1940, 1960, 1980, 2000, 2020)

Generate percentage changes across start end end points of each bin, for bins of varying length. For both raw and interpolated data.

# E. Generate cuts -----
it_avg_type <- 1
it_cut_type <- 1
for (it_avg_type in c(1, 2)) {
  
  if (it_avg_type == 1) {
    svr_chg_var <- "pchg_yr1"
    # common var across all spans
    svr_chg_var_new <- "pchg"
    svr_var_val <- "value"
  } else if (it_avg_type == 2) {
    svr_chg_var <- "pchg_yr1_interp1"
    # common var across all spans
    svr_chg_var_new <- "pchg_interp1"
    svr_var_val <- "value_interp1"
  }
  
  # Add to stack the annual results
  ppts_easia_weuro_pchg <- ppts_easia_weuro_long %>%
    select(location_code, location_level,
           variable, year, one_of(svr_chg_var, svr_var_val)) %>%
    filter(!is.na(!!sym(svr_chg_var)) | !is.na(!!sym(svr_var_val))) %>%
    rename(!!sym(svr_chg_var_new) := !!sym(svr_chg_var),
           year_bins = year) %>%
    mutate(year_bins = as.factor(year_bins)) %>%
    mutate(year_bins_type = "1940t2020i01") %>%
    ungroup()
  
  # Loop over cut types    
  for (st_cuts_name in names(ls_ar_it_cuts)) {
    print(glue::glue(
      "cutTypeName={st_cuts_name}:\n",
      "bins={ls_ar_it_cuts[st_cuts_name]}"))
    
    # temp dataframe
    ppts_easia_weuro_long_cut <- ppts_easia_weuro_long %>%
      select(location_code, location_level,
             variable, year,
             one_of(svr_chg_var, svr_var_val))
    
    # E.1 Cut types
    ar_it_cuts <- ls_ar_it_cuts[[st_cuts_name]]
    it_gap <- ar_it_cuts[2] - ar_it_cuts[1]
    ar_it_end_seg <- ar_it_cuts[2:length(ar_it_cuts)]
    ar_it_start_seg <- ar_it_end_seg - it_gap + 1
    ar_st_lab <- paste0(ar_it_start_seg, "-", ar_it_end_seg)
    
    # E.2 Generate new year groupings, consider only full-segments
    # consider only sub-segments with observations in all years
    ppts_easia_weuro_long_cut <- ppts_easia_weuro_long_cut %>%
      mutate(year_bins = cut(year,
                             breaks = ar_it_cuts,
                             labels = ar_st_lab,
                             right = TRUE)) %>%
      group_by(location_code, location_level,
               variable, year_bins) %>%
      mutate(val_n_in_bin = sum(!is.na(!!sym(svr_chg_var))))
    # filter(val_n_in_bin == it_gap)
    
    # E.3 cumulative product
    ppts_easia_weuro_long_cut <- ppts_easia_weuro_long_cut %>%
      arrange(location_code, location_level, variable, year_bins, year) %>%
      group_by(location_code, location_level, variable, year_bins) %>%
      mutate(!!sym(svr_chg_var_new) := cumprod(1 + !!sym(svr_chg_var)) - 1) %>% 
      mutate(!!sym(svr_chg_var_new) := 
        case_when(val_n_in_bin == it_gap ~ !!sym(svr_chg_var_new),
                  TRUE ~ NA))

    # View(ppts_easia_weuro_long_cut)
    
    # E.4, slices last row
    ppts_easia_weuro_long_cut <- ppts_easia_weuro_long_cut %>%
      slice(n()) %>%
      select(location_code, location_level,
             variable, year_bins, one_of(svr_chg_var_new, svr_var_val)) %>%
      ungroup() %>%
      mutate(year_bins_type = st_cuts_name) %>%
      filter(!is.na(!!sym(svr_chg_var_new)) | !is.na(!!sym(svr_var_val))) %>% 
      drop_na(year_bins)
    
    # E.5 Stack
    ppts_easia_weuro_pchg <- bind_rows(
      ppts_easia_weuro_pchg, ppts_easia_weuro_long_cut)
  }
  
  # export
  if (it_avg_type == 1) {
    ppts_easia_weuro_pchg_raw <- ppts_easia_weuro_pchg
  } else if (it_avg_type == 2) {
    ppts_easia_weuro_pchg_interp1 <- ppts_easia_weuro_pchg
  }
  
  # # Print results
  # print(kable(ppts_easia_weuro_long %>% 
  #         filter(location_code == 'AFG' & variable == 'student'), 
  #        caption= paste0("breaks=", st_cuts_name, ", variable=", svr_chg_var_new)))
}
#> cutTypeName=1920t2020i05:
#> bins=c(1920, 1925, 1930, 1935, 1940, 1945, 1950, 1955, 1960, 1965, 1970, 1975, 1980, 1985, 1990, 1995, 2000, 2005, 2010, 2015, 2020)
#> cutTypeName=1920t2020i10:
#> bins=c(1920, 1930, 1940, 1950, 1960, 1970, 1980, 1990, 2000, 2010, 2020)
#> cutTypeName=1925t2015i15:
#> bins=c(1925, 1940, 1955, 1970, 1985, 2000, 2015)
#> cutTypeName=1920t2020i20:
#> bins=c(1920, 1940, 1960, 1980, 2000, 2020)
#> cutTypeName=1920t2020i05:
#> bins=c(1920, 1925, 1930, 1935, 1940, 1945, 1950, 1955, 1960, 1965, 1970, 1975, 1980, 1985, 1990, 1995, 2000, 2005, 2010, 2015, 2020)
#> cutTypeName=1920t2020i10:
#> bins=c(1920, 1930, 1940, 1950, 1960, 1970, 1980, 1990, 2000, 2010, 2020)
#> cutTypeName=1925t2015i15:
#> bins=c(1925, 1940, 1955, 1970, 1985, 2000, 2015)
#> cutTypeName=1920t2020i20:
#> bins=c(1920, 1940, 1960, 1980, 2000, 2020)

Merge raw and interpolated results together, again

After generating percentage changes over different spans, merge again.

# D.3 merge together
# full_join same as left_join, 
ppts_easia_weuro_world_pchg <- ppts_easia_weuro_pchg_interp1 %>%
  full_join(ppts_easia_weuro_pchg_raw,
            by = (c('location_code' = 'location_code',
                    'location_level' = 'location_level',
                    'variable' = 'variable', 
                    'year_bins_type' = 'year_bins_type',
                    'year_bins' = 'year_bins'
            ))) %>% 
  mutate(variable = as.factor(variable), 
         year_bins_type = as.factor(year_bins_type), 
         # year_bins as string to allow for correct sorting
         year_bins = as.character(year_bins)) %>%
  select(location_code, location_level, 
         variable, 
         year_bins_type, year_bins, 
         pchg, pchg_interp1, 
         value, value_interp1) 
# %>% 
  # value has missing data at the start year of each perentage calculation
  # value_interp1 is just raw data but complete
  # select(-value) %>% rename(value = value_interp1)

str(ppts_easia_weuro_world_pchg)
#> tibble [82,685 × 9] (S3: tbl_df/tbl/data.frame)
#>  $ location_code : Factor w/ 286 levels "ABW","AFE","AFG",..: 1 1 1 1 1 1 1 1 1 1 ...
#>  $ location_level: Factor w/ 4 levels "country","multicountry",..: 1 1 1 1 1 1 1 1 1 1 ...
#>  $ variable      : Factor w/ 5 levels "gdp","school",..: 1 1 1 1 1 1 1 1 1 1 ...
#>  $ year_bins_type: Factor w/ 5 levels "1920t2020i05",..: 5 5 5 5 5 5 5 5 5 5 ...
#>  $ year_bins     : chr [1:82685] "1981" "1982" "1983" "1984" ...
#>  $ pchg          : num [1:82685] NA NA NA NA NA ...
#>  $ pchg_interp1  : num [1:82685] 0.176 0.176 0.176 0.176 0.176 ...
#>  $ value         : num [1:82685] NA NA NA NA NA ...
#>  $ value_interp1 : num [1:82685] 7662 9010 10596 12460 14653 ...
# print(ppts_easia_weuro_world_pchg[1:50,])

Sort and Display

We sort variables for more clearly organized output file.

# Arrange results 
ppts_easia_weuro_world_pchg <- ppts_easia_weuro_world_pchg %>%
  arrange(year_bins_type, 
          location_level, location_code, 
          variable, year_bins)

Afghanistan as Example

We print results from year year group aggregation. The Afghanistan example demonstrates that we are able to compute more interval-percentage changes after interpolation, given filled out values at key years. We use student count as example.

Note that the “value” column shows value at the last year in the interval, if this was observed.

# print
for (st_cuts_name in names(ls_ar_it_cuts)) {
  print(
    kable(ppts_easia_weuro_world_pchg %>% 
            filter(location_code == 'AFG' & 
                     variable == 'student' &
                     year_bins_type == st_cuts_name) %>% 
            select(-location_level, -year_bins_type),
          caption = paste0("Afghanistan example, aggregate=", st_cuts_name)))
}

Afghanistan example, aggregate=1920t2020i05

location_code	variable	year_bins	pchg	pchg_interp1	value	value_interp1
AFG	student	1961-1965	NA	NA	NA	404712.6
AFG	student	1966-1970	NA	0.3359727	540685	540685.0
AFG	student	1971-1975	0.2804905	0.2804905	692342	692342.0
AFG	student	1976-1980	NA	0.3859956	959583	959583.0
AFG	student	1981-1985	NA	-0.5006685	479150	479150.0
AFG	student	1986-1990	NA	0.2992028	622513	622513.0
AFG	student	1991-1995	NA	1.1079030	1312197	1312197.0
AFG	student	1996-2000	NA	-0.4289272	749360	749360.0
AFG	student	2001-2005	4.7633434	4.7633434	4318819	4318819.0
AFG	student	2006-2010	0.2224004	0.2224004	5279326	5279326.0
AFG	student	2011-2015	0.1742652	0.1742652	6199329	6199329.0
AFG	student	2016-2020	NA	0.1322113	NA	7018950.2

Afghanistan example, aggregate=1920t2020i10

location_code	variable	year_bins	pchg	pchg_interp1	value	value_interp1
AFG	student	1961-1970	NA	NA	540685	540685
AFG	student	1971-1980	NA	0.7747542	959583	959583
AFG	student	1981-1990	NA	-0.3512672	622513	622513
AFG	student	1991-2000	NA	0.2037660	749360	749360
AFG	student	2001-2010	6.045113	6.0451132	5279326	5279326
AFG	student	2011-2020	NA	0.3295164	NA	7018950

Afghanistan example, aggregate=1925t2015i15

location_code	variable	year_bins	pchg	pchg_interp1	value	value_interp1
AFG	student	1956-1970	NA	NA	540685	540685
AFG	student	1971-1985	NA	-0.1138093	479150	479150
AFG	student	1986-2000	NA	0.5639361	749360	749360
AFG	student	2001-2015	7.272831	7.2728315	6199329	6199329

Afghanistan example, aggregate=1920t2020i20

location_code	variable	year_bins	pchg	pchg_interp1	value	value_interp1
AFG	student	1961-1980	NA	NA	959583	959583
AFG	student	1981-2000	NA	-0.2190775	749360	749360
AFG	student	2001-2020	NA	8.3665931	NA	7018950

Austria as Example

Below we illustrate with Austria as example, showing data for school count. We are able to fill out some initial decades/years a little bit more with interpolation.

# print
for (st_cuts_name in names(ls_ar_it_cuts)) {
  print(
    kable(ppts_easia_weuro_world_pchg %>% 
            filter(location_code == 'AUT' & 
                     variable == 'school' &
                     year_bins_type == st_cuts_name) %>%
            select(-location_level, -year_bins_type), 
          caption = paste0("Austria example, aggregate=", st_cuts_name)))
}

Austria example, aggregate=1920t2020i05

location_code	variable	year_bins	pchg	pchg_interp1	value	value_interp1
AUT	school	1921-1925	NA	NA	4697	4697.000
AUT	school	1926-1930	NA	0.0004258	4699	4699.000
AUT	school	1931-1935	NA	-0.0053203	4674	4674.000
AUT	school	1936-1940	NA	-0.0415880	NA	4479.618
AUT	school	1941-1945	NA	-0.0415880	NA	4293.319
AUT	school	1946-1950	NA	0.0248480	4400	4400.000
AUT	school	1951-1955	0.0061364	0.0061364	4427	4427.000
AUT	school	1956-1960	-0.0076801	-0.0076801	4393	4393.000
AUT	school	1961-1965	-0.0223082	-0.0223082	4295	4295.000
AUT	school	1966-1970	-0.0749709	-0.0749709	3973	3973.000
AUT	school	1971-1975	-0.0964007	-0.0964007	3590	3590.000
AUT	school	1976-1980	-0.0389972	-0.0389972	3450	3450.000
AUT	school	1981-1985	-0.0113043	-0.0113043	3411	3411.000
AUT	school	1986-1990	-0.0073292	-0.0073292	3386	3386.000
AUT	school	1991-1995	-0.0008860	-0.0008860	3383	3383.000
AUT	school	1996-2000	-0.0067987	-0.0067987	3360	3360.000
AUT	school	2001-2005	-0.0190476	-0.0190476	3296	3296.000
AUT	school	2006-2010	-0.0379248	-0.0379248	3171	3171.000
AUT	school	2011-2015	-0.0416272	-0.0416272	3039	3039.000
AUT	school	2016-2020	-0.0082264	-0.0082264	3014	3014.000

Austria example, aggregate=1920t2020i10

location_code	variable	year_bins	pchg	pchg_interp1	value	value_interp1
AUT	school	1921-1930	NA	NA	4699	4699.000
AUT	school	1931-1940	NA	-0.0466870	NA	4479.618
AUT	school	1941-1950	NA	-0.0177733	4400	4400.000
AUT	school	1951-1960	-0.0015909	-0.0015909	4393	4393.000
AUT	school	1961-1970	-0.0956066	-0.0956066	3973	3973.000
AUT	school	1971-1980	-0.1316386	-0.1316386	3450	3450.000
AUT	school	1981-1990	-0.0185507	-0.0185507	3386	3386.000
AUT	school	1991-2000	-0.0076787	-0.0076787	3360	3360.000
AUT	school	2001-2010	-0.0562500	-0.0562500	3171	3171.000
AUT	school	2011-2020	-0.0495112	-0.0495112	3014	3014.000

Austria example, aggregate=1925t2015i15

location_code	variable	year_bins	pchg	pchg_interp1	value	value_interp1
AUT	school	1926-1940	NA	-0.0462811	NA	4479.618
AUT	school	1941-1955	NA	-0.0117460	4427	4427.000
AUT	school	1956-1970	-0.1025525	-0.1025525	3973	3973.000
AUT	school	1971-1985	-0.1414548	-0.1414548	3411	3411.000
AUT	school	1986-2000	-0.0149516	-0.0149516	3360	3360.000
AUT	school	2001-2015	-0.0955357	-0.0955357	3039	3039.000

Austria example, aggregate=1920t2020i20

location_code	variable	year_bins	pchg	pchg_interp1	value	value_interp1
AUT	school	1921-1940	NA	NA	NA	4479.618
AUT	school	1941-1960	NA	-0.0193360	4393	4393.000
AUT	school	1961-1980	-0.2146597	-0.2146597	3450	3450.000
AUT	school	1981-2000	-0.0260870	-0.0260870	3360	3360.000
AUT	school	2001-2020	-0.1029762	-0.1029762	3014	3014.000

Switerland as Example

Below we illustrate with Germany as example, where we filled in values.

# print
for (st_cuts_name in names(ls_ar_it_cuts)) {
  print(
    kable(ppts_easia_weuro_world_pchg %>% 
            filter(location_code == 'CHE' & 
                     variable == 'teacher' &
                     year_bins_type == st_cuts_name) %>%
            select(-location_level, -year_bins_type), 
          caption = paste0("Germany example, aggregate=", st_cuts_name)))
}

Germany example, aggregate=1920t2020i05

location_code	variable	year_bins	pchg	pchg_interp1	value	value_interp1
CHE	teacher	1921-1925	NA	-0.0154381	NA	13289.02
CHE	teacher	1926-1930	NA	-0.0012891	NA	13271.88
CHE	teacher	1931-1935	NA	0.0229897	13577	13577.00
CHE	teacher	1936-1940	NA	-0.0031722	NA	13533.93
CHE	teacher	1941-1945	NA	0.0117872	NA	13693.46
CHE	teacher	1946-1950	NA	0.0470723	NA	14338.04
CHE	teacher	1951-1955	NA	0.1171946	NA	16018.38
CHE	teacher	1956-1960	NA	0.0893236	NA	17449.20
CHE	teacher	1961-1965	NA	0.0936521	NA	19083.36
CHE	teacher	1966-1970	NA	0.0975457	NA	20944.85
CHE	teacher	1971-1975	NA	0.0975457	NA	22987.94
CHE	teacher	1976-1980	NA	0.0975457	NA	25230.31
CHE	teacher	1981-1985	NA	0.0975457	NA	27691.42
CHE	teacher	1986-1990	NA	0.0975457	NA	30392.60
CHE	teacher	1991-1995	NA	0.0975457	NA	33357.26
CHE	teacher	1996-2000	NA	0.0975457	NA	36611.12
CHE	teacher	2001-2005	NA	0.0975457	NA	40182.38
CHE	teacher	2006-2010	NA	0.0975457	44102	44102.00
CHE	teacher	2011-2015	0.1018775	0.1018775	48595	48595.00
CHE	teacher	2016-2020	0.1272147	0.1272147	54777	54777.00

Germany example, aggregate=1920t2020i10

location_code	variable	year_bins	pchg	pchg_interp1	value	value_interp1
CHE	teacher	1921-1930	NA	-0.0167073	NA	13271.88
CHE	teacher	1931-1940	NA	0.0197445	NA	13533.93
CHE	teacher	1941-1950	NA	0.0594144	NA	14338.04
CHE	teacher	1951-1960	NA	0.2169864	NA	17449.20
CHE	teacher	1961-1970	NA	0.2003332	NA	20944.85
CHE	teacher	1971-1980	NA	0.2046066	NA	25230.31
CHE	teacher	1981-1990	NA	0.2046066	NA	30392.60
CHE	teacher	1991-2000	NA	0.2046066	NA	36611.12
CHE	teacher	2001-2010	NA	0.2046066	44102	44102.00
CHE	teacher	2011-2020	0.2420525	0.2420525	54777	54777.00

Germany example, aggregate=1925t2015i15

location_code	variable	year_bins	pchg	pchg_interp1	value	value_interp1
CHE	teacher	1926-1940	NA	0.0184299	NA	13533.93
CHE	teacher	1941-1955	NA	0.1835720	NA	16018.38
CHE	teacher	1956-1970	NA	0.3075513	NA	20944.85
CHE	teacher	1971-1985	NA	0.3221108	NA	27691.42
CHE	teacher	1986-2000	NA	0.3221108	NA	36611.12
CHE	teacher	2001-2015	NA	0.3273289	48595	48595.00

Germany example, aggregate=1920t2020i20

location_code	variable	year_bins	pchg	pchg_interp1	value	value_interp1
CHE	teacher	1921-1940	NA	0.0027073	NA	13533.93
CHE	teacher	1941-1960	NA	0.2892929	NA	17449.20
CHE	teacher	1961-1980	NA	0.4459293	NA	25230.31
CHE	teacher	1981-2000	NA	0.4510770	NA	36611.12
CHE	teacher	2001-2020	NA	0.4961846	54777	54777.00

Germany as Example

Below we illustrate with Germany as example, where we filled in values.

# print
for (st_cuts_name in names(ls_ar_it_cuts)) {
  print(
    kable(ppts_easia_weuro_world_pchg %>% 
            filter(location_code == 'DEU' & 
                     variable == 'school' &
                     year_bins_type == st_cuts_name) %>%
            select(-location_level, -year_bins_type), 
          caption = paste0("Germany example, aggregate=", st_cuts_name)))
}

Germany example, aggregate=1920t2020i05

location_code	variable	year_bins	pchg	pchg_interp1	value	value_interp1
DEU	school	1986-1990	NA	NA	NA	18001.15
DEU	school	1991-1995	NA	-0.0050636	17910	17910.00
DEU	school	1996-2000	-0.0354551	-0.0354551	17275	17275.00
DEU	school	2001-2005	-0.0266860	-0.0266860	16814	16814.00
DEU	school	2006-2010	-0.0311645	-0.0311645	16290	16290.00
DEU	school	2011-2015	-0.0495396	-0.0495396	15483	15483.00
DEU	school	2016-2020	-0.0023251	-0.0023251	15447	15447.00

Germany example, aggregate=1920t2020i10

location_code	variable	year_bins	pchg	pchg_interp1	value	value_interp1
DEU	school	1981-1990	NA	NA	NA	18001.15
DEU	school	1991-2000	NA	-0.0403391	17275	17275.00
DEU	school	2001-2010	-0.0570188	-0.0570188	16290	16290.00
DEU	school	2011-2020	-0.0517495	-0.0517495	15447	15447.00

Germany example, aggregate=1925t2015i15

location_code	variable	year_bins	pchg	pchg_interp1	value	value_interp1
DEU	school	1986-2000	NA	NA	17275	17275
DEU	school	2001-2015	-0.1037337	-0.1037337	15483	15483

Germany example, aggregate=1920t2020i20

location_code	variable	year_bins	pchg	pchg_interp1	value	value_interp1
DEU	school	1981-2000	NA	NA	17275	17275
DEU	school	2001-2020	-0.1058177	-0.1058177	15447	15447

Store to file

Finally, we save results to file in the data folder.

# Write to CSV and write to rda
if (bl_resave_to_data) {
  write_csv(ppts_easia_weuro_world_pchg, "../data/ppts_easia_weuro_world_pchg.csv", na="")
  usethis::use_data(ppts_easia_weuro_world_pchg, overwrite = TRUE)
}