The CBRT package includes functions for finding, and downloading data from the Central Bank of the Republic of Turkey’s database.
The CBRT database covers more than 40,000 time series variables.
You can install the package from the source by using the following command (the CBRT package depends on the data.table package.):
{r, eval = F} install.packages("http://users.metu.edu.tr/etaymaz/files/CBRT_0.1.1.tar.gz", repos = NULL, type = "source")
You may also install it from GitHub. Install the the
devtools package if not
installed, then
{r, eval = F} library(devtools) install_github("etaymaz/CBRT")
Please kindly note that you need a key to download data from the CBRT’s database. To get the key, register at the CBRT’s Electronic Data Delivery System. Registration is free of charge and takes only a few minutes.
If you create an object called myCBRTkey in R session, you do not need to define it for downloading:
myCBRTKey <-
your-key
All data series (variables) are classified into data groups, and data groups into data categories. There are 44 data categories (including the archieved ones), 496 data groups, and 40,826 data series.
To find variables, use the searchCBRT
function:
{r, p0, eval = FALSE} searchCBRT(c("production", "labor", "labour")) searchCBRT(c("production", "labor", "labour"), field = "series") searchCBRT(c("production", "labor", "labour"), tags = TRUE)
The package contains the lists of all data categories, data groups, and data series, as of 26 January 2019. You can update the lists by the following commands:
{r, p1, eval = FALSE} allCBRTCategories <- getAllCategoriesInfo() allCBRTGroups <- getAllGroupsInfo() allCBRTSeries <- getAllSeriesInfo()
After identifying the data group or data series, you can get some
information about the data by showGroupInfo
function:
{r, p2, eval = FALSE} showGroupInfo("bie_apifon")
If you want to get only names of series in a data group, use the following command:
{r, p3, eval = FALSE} showSeriesNames("bie_apifon")
You
can download either one or more data series you specified, or all data
series in a data group.
To download individual data series, use the
getDataSeries
function:
{r, p4, eval = FALSE} mySeries <- getDataSeries("TP.D1TOP") mySeries <- getDataSeries(c("TP.D1TOP", "TP.D2HAZ", "TP.D4TCMB")) mySeries <- getDataSeries(c("TP.D1TOP", "TP.D2HAZ", "TP.D4TCMB", startDate="01-01-2010"))
To download all data series in a group, use the
getDataGroup
function:
{r, p5, eval = FALSE} myData <- getDataGroup("bie_dbafod")
The freq
parameter defines the frequency of the data. If
you do not define any frequency, the default frequency will be used.
The aggType
paremeter defines the method to be used to
aggregate data series from high frequency to low frequency (for example,
weekly data to monthly data). If no aggregation method is defined, the
default will be used. (For the default values, use the
showGroupInfo
function.)
For example, if you define monthly frequency for weekly data, and
“sum” as the aggregation method, then the monthly totals will be
returned. Since a data group includes more than one series, the
getDataGroup
function does not have any
aggType
parameter, and it aggregates data series by using
their default aggregation method.
The following frequencies are defined (from high frequency to low frequency):
1
Day2
Work day3
Week4
Biweekly5
Month6
Quarter7
Six months8
YearThe following aggregation methods are available:
avg
Average valuefirst
First observationlast
Last observationmax
Maximum valuemin
Minimum valuesum
SumThe myData object is in data.table and
data.frame classes, and it includes a
time variable, and data series. The
time variable will be either in date
or
numeric
format depending on its frequency.
I would appreciate your comments, suggestions, and bug reports. Please contact me by e-mail.