Disclaimer
This software is in no way affiliated, endorsed, or approved by the CBRT. It comes with absolutely no warranty. Also see the Disclaimer on CBRT webpage since it mentions; “Information published in this site may be quoted by specific reference thereto, but the use of such information for commercial purposes shall be subject to prior written permission of the CBRT.”
Overview
cbRt is an R interface to The Electronic Data Delivery System (EDDS) of Central Bank of the Republic of Turkey (CBRT). “The system provides a rich range of economic data and information to support economic education and foster economic research.”[1]
Similar Packages
- See etaymaz/CBRT
Installation
You can install cbRt from github with:
# install.packages("devtools")
devtools::install_github("emraher/cbRt")Usage and Arguments
There are two functions in the package.
cbrt_meta retrieves all information for all series.
series_meta_info <- cbrt_meta(token = NULL)For token argument see the definition below. This returns a tibble. One can search this tibble to find series IDs and other information on series.
The other function is cbrt_get which retrieves data from EDDS.
cbrt_get(series, start_date, end_date, formulas, token = NULL, nd = TRUE,
as = c("tibble", "tsibble", "data.frame", "data.table"))Following are the arguments used in the functions. Descriptions are taken from EDDS Web Service Usage Guide.
Note that, series, start_date, end_date, and token are all required arguments.
series
series argument can be obtained from CBRT webpage. Search CBRT webpage in order to find out the series code and use it in the series argument. For example, TP.DK.USD.A is the code for (USDTRY) US Dollar (Buying) Exchange Rate.
One can also use cbrt_meta function to get all information for all series and search in the resulting tibble.
Argument can take multiple values.
start_date
start_date argument is the series start date as dd-mm-yyyy format. This argument takes a single value.
end_date
end_date argument is the series end date as dd-mm-yyyy format. This argument takes a single value.
formulas
formulas argument is the formula applied to series. Available formulas are;
- 0: Level
- 1: Percentage Change
- 2: Difference
- 3: Year-to-year Percent Change
- 4: Year-to-year Differences
- 5: Percentage Change Compared to End-of-Previous Year
- 6: Difference Compared to End-of-Previous Year
- 7: Moving Average
- 8: Moving Sum
If this parameter is not supplied by the user, the level formula parameter is applied for the relevant series. If retrieving multiple series, argument should take multiple values.
token
token argument is the required API key. Follow the instructions on CBRT webpage to obtain the user specific API key. This argument takes a single value.
You can set your token with
or you can add it to your .Renviron file as
Examples
Single Series
Following example retrieves (USDTRY) US Dollar (Buying) Exchange Rate.
library(cbRt)
series <- "TP.DK.USD.A"
start_date <- "01-01-2017"
end_date <- "01-01-2018"
token <- Sys.getenv("EVDS_TOKEN")
(usd_try <- cbrt_get(series, start_date, end_date, token = token))
#> # A tibble: 366 × 2
#> Date TP_DK_USD_A
#> <date> <dbl>
#> 1 2017-01-01 NA
#> 2 2017-01-02 3.52
#> 3 2017-01-03 3.53
#> 4 2017-01-04 3.57
#> 5 2017-01-05 3.58
#> 6 2017-01-06 3.59
#> 7 2017-01-07 NA
#> 8 2017-01-08 NA
#> 9 2017-01-09 3.61
#> 10 2017-01-10 3.70
#> # ℹ 356 more rowsMultiple Series
Following example retrieves multiple different series with different frequencies.
EDDS API converts series to a common frequency if they are requested together and no frequency argument is given.
For example, if you request one yearly and one monthly series, API will return both series as yearly values.
Package, on the other hand, sends independent queries for each series and joins them together without changing the frequency.
I opted not to include freq argument to function call.
The example below also shows the usage of the formulas argument.
start_date <- "01-01-2017"
end_date <- "01-01-2020"
series <- c("TP.AB.B1", "TP.AB.C2", "TP.BKEA.S001", "TP.KB.O06.TRL")
formulas <- c(2, 3, 7, 8)
token <- Sys.getenv("EVDS_TOKEN") # Get token from .Renviron
# data(cbrt_meta_data)
# cbrt_meta_data %>%
# dplyr::filter(SERIE_CODE %in% series) %>%
# dplyr::select(SERIE_CODE, FREQUENCY_STR)
#
# SERIE_CODE FREQUENCY_STR
# 1 TP.AB.C2 HAFTALIK(CUMA)
# 2 TP.AB.B1 AYLIK
# 3 TP.BKEA.S001 ÜÇ AYLIK
# 4 TP.KB.O06.TRL YILLIK
# EDDS API returns following
url <- paste0("https://evds2.tcmb.gov.tr/service/evds/series=TP.AB.B1-TP.AB.C2-TP.BKEA.S001-TP.KB.O06.TRL&startDate=01-01-2017&endDate=01-01-2020&type=json", "&formulas=2-3-7-8")
(edds_dat <- cbRt:::cbrt_geturl(url, token = token) %>%
httr::content(as = "text", encoding = "UTF-8") %>%
jsonlite::fromJSON() %>%
.[["items"]] %>%
tibble::as_tibble())
#> # A tibble: 4 × 10
#> Tarih `TP_AB_B1-2` `TP_AB_C2-3` `TP_BKEA_S001-7` `TP_KB_O06_TRL-8`
#> <chr> <chr> <chr> <chr> <chr>
#> 1 2017 9485.7 -8.625746876697447 0 343279699.6
#> 2 2018 -3407 -14.421590773986445 -20.3 399316709.55
#> 3 2019 6957 12.864684634620726 57.1 471538564.43
#> 4 2020 16466 -38.49704579025111 1.2 144592850
#> # ℹ 5 more variables: UNIXTIME <df[,1]>, TP_AB_B1 <lgl>, TP_AB_C2 <lgl>,
#> # TP_BKEA_S001 <lgl>, TP_KB_O06_TRL <lgl>NOTE that API call returns empty columns. cbrt_get function also removes them.
(dat <- cbrt_get(series, start_date, end_date, as = "tsibble"))
#> # A tsibble: 188 x 5 [1D]
#> Date TP_AB_B1 TP_AB_C2 TP_BKEA_S001 TP_KB_O06_TRL
#> <date> <dbl> <dbl> <dbl> <dbl>
#> 1 2017-01-01 15933 NA 0 343279700.
#> 2 2017-01-06 NA 96933 NA NA
#> 3 2017-01-13 NA 95292 NA NA
#> 4 2017-01-20 NA 92296 NA NA
#> 5 2017-01-27 NA 92517 NA NA
#> 6 2017-02-01 16648 NA NA NA
#> 7 2017-02-03 NA 91522 NA NA
#> 8 2017-02-10 NA 92798 NA NA
#> 9 2017-02-17 NA 89049 NA NA
#> 10 2017-02-24 NA 91088 NA NA
#> # ℹ 178 more rowsSeries with Location Data
Following are examples for data with location attributes. Series is Hedonic House Price Index (HHPI)(2010=100)(CBRT)(Monthly). TP.HKFE02 and TP.HKFE03 are price indices codes for Istanbul and Ankara, respectively. This series can be used to draw maps. At this time there is no support in the package though.
# Single Series
series <- "TP.HKFE02"
(hhpi <- cbrt_get(series, start_date, end_date))
#> # A tibble: 37 × 2
#> Date TP_HKFE02
#> <date> <dbl>
#> 1 2017-01-01 97.5
#> 2 2017-02-01 98.6
#> 3 2017-03-01 99.4
#> 4 2017-04-01 99.8
#> 5 2017-05-01 100.
#> 6 2017-06-01 100.
#> 7 2017-07-01 100.
#> 8 2017-08-01 100.
#> 9 2017-09-01 100.
#> 10 2017-10-01 101.
#> # ℹ 27 more rows
# Multiple Series
series <- c("TP.HKFE02", "TP.HKFE03")
(hhpi <- cbrt_get(series, start_date, end_date))
#> # A tibble: 37 × 3
#> Date TP_HKFE02 TP_HKFE03
#> <date> <dbl> <dbl>
#> 1 2017-01-01 97.5 96.4
#> 2 2017-02-01 98.6 97.3
#> 3 2017-03-01 99.4 98.6
#> 4 2017-04-01 99.8 99.2
#> 5 2017-05-01 100. 100.
#> 6 2017-06-01 100. 100.
#> 7 2017-07-01 100. 100.
#> 8 2017-08-01 100. 99.9
#> 9 2017-09-01 100. 101.
#> 10 2017-10-01 101. 102.
#> # ℹ 27 more rows