A package that implements an R client for the ONS web API.
# If the devtools package is not yet installed, run:
# install.packages ("devtools")
devtools::install_github("https://github.com/hamstr147/onsRapi")
library(onsRapi)
A user can see what datasets are available using the list_datasets function. An optional include_desc argument can be set to TRUE to return the dataset IDs alongside full descriptions in a data frame.
datasets <- list_datasets()
head(datasets)
id
1 cpih01
2 mid-year-pop-est
3 ashe-table-7-hours
4 ashe-table-7-earnings
5 ashe-table-8-hours
6 ashe-table-8-earnings
7 opss-rates
8 opss-membership
9 wellbeing-year-ending
10 wellbeing-local-authority
Suppose we want to look at the cpih01 data set (which shows inflation as measured using CPI). Firstly, we can see what dimensions we can use to filter the data using the get_code_lists function.
cpi_codes <- get_code_lists("cpih01")
names(cpi_codes)
[1] "time" "geography" "aggregate"
So we can filter this dataset using any of these dimensions. To see what values we can use to filter the data, we can inspect the individual elements of cpi_codes, which is a list. I'm not sure what 'aggregate' is so let's take a look at that.
head(cpi_codes$aggregate)
code label
1 cpih1dim1S90401 09.4.1 Recreational and sporting services
2 cpih1dim1S90501 09.5.1 Books
3 cpih1dim1S90402 09.4.2 Cultural services
4 cpih1dim1T90000 09 Recreation and culture
5 cpih1dim1T80000 08 Communication
6 cpih1dim1T10000 01 Food and non-alcoholic beverages
I'm interested in inflation in book prices, so let's download that data. We can use the 'code' value in the code list data frames to filter the data - we can provide as many or few filters as we want, and if we don't provide any at all, I get the full data set!
In this case, I'll use the cpih1dim1S90501 value to filter on aggregate, and I won't filter on geography or time, so I will get results for book-related CPI for all times and geographies available. All I need to do is pass the id of the dataset ("cpih01") and a filter (aggregate = "cpih1dim1S90501") to the get_dataset function:
book_cpi <- get_dataset("cpih01", aggregate = "cpih1dim1S90501")
Getting data set...
[100%] Downloaded 28740 bytes...
Done!
head(book_cpi)
V4_0 mmm-yy time uk-only geography cpih1dim1aggid aggregate
1: 85.0 Jan-12 Jan-12 K02000001 United Kingdom cpih1dim1S90501 09.5.1 Books
2: 93.9 Oct-11 Oct-11 K02000001 United Kingdom cpih1dim1S90501 09.5.1 Books
3: 61.1 May-94 May-94 K02000001 United Kingdom cpih1dim1S90501 09.5.1 Books
4: 80.5 Feb-05 Feb-05 K02000001 United Kingdom cpih1dim1S90501 09.5.1 Books
5: 68.9 May-98 May-98 K02000001 United Kingdom cpih1dim1S90501 09.5.1 Books
6: 49.3 Apr-90 Apr-90 K02000001 United Kingdom cpih1dim1S90501 09.5.1 Books
And there you go - book inflation in the UK all the way back to the 90s.
This package is an R wrapper for the filtering functionality as outlined on the ONS website here. It sends options for each dimension via POST to create a filtered dataset, which can then be retrieved through a GET on the filter output. This provides a more general way of retrieving datasets than just using GET requests - the current ONS API only support the use of one wildcard in each query. With filtering, we can get anything from the full data set to a single observation using a single method.