In cormac85/datakindr: Tools And Utilities for DataKind

library(ggplot2)
library(knitr)
library(datakindr)

opts_chunk$set(fig.width=6, fig.height=4)
opts_chunk$set(echo = F, warning = F, error = F, message = F)

What Is It For?

Datakindr provides tools, utilities, data and more for use by the DataKind Dublin team. It includes a convenient interface to CSO.ie's Statbank, a ggplot2 theme and colour palettes for convenient & consistent visualisations.

The package is in the early stages of development but it is intended that this package will become a kind of "Swiss Army Knife". It should mean easy and frictionless solving of commonly encountered problems for Datakind volunteers (e.g., querying the CSO's Statbank).

What Does It Do?

Currently there are two main tasks that datakindr makes easier:

Querying the CSO's Statbank API.
Making consistent & beautiful visualisations.

CSO and Statbank

There are two functions available for this task:

search_statbank_dataset() Gets a list of datasets from a search query.
get_cso_dataset() Gets the actual data from Statbank.

Visualisations

There are two sets of data objects here for helping volunteers create consistent plots and other visualisations.

dk_theme A ggplot2 object that will help you effortlessly create visualisations like this:

ggplot(data.frame( y = runif(100, min = 0, max = 20) +
                     seq(1, 100, 1),
                   x = seq(1, 100, 1),
                   z = rep_len(c("a", "b", "c", "d"), 100)),
       aes(x, y, colour = z, fill = z)) +
  geom_point() +
  facet_wrap(~z, nrow = 1) +
  geom_smooth(se = TRUE) +
  labs(title = "Some Data (2011)",
       x = "Range", y = "Value") +
  dk_theme

dk_palette A palette of 2 primary, 2 secondary and 2 background colours. The primary and secondary colours can be seen above.
- dk_sequential_palette An extended version of the above palette intended for use with sequential data.
- dk_non_sequential_palette Another extended palette for (obviously!) non-sequential data.

How Do I Use It?

I'll let the code do the talking here.

Theme Demo

# install.packages("devtools")
# devtools::install_github("DataKind-DUB/datakindr")
# Development version: devtools::install_github("cormac85/datakindr")
library(datakindr)

data_example <- data.frame( x = c('Primary 1', 'Primary 2',
                                  'Secondary 1', 'Secondary 2'),
                            y = runif(4, 10, 100))

ggplot(data_example, aes(x,y, fill = x)) +
  geom_bar(stat = "identity") +
  labs(title = "Some Data (2011)",
       x = "Range", y = "Value") +
  dk_theme # <- Here's the theme!

Statbank Demo

library(dplyr)
dataset_names <- search_statbank_datasets("garda")
dataset_names[1,1]
dataset_names[1,2]

industry_population <- get_cso_dataset(dataset_names$dataset_code[1])

industry_population %>% 
  filter(grepl("Both", Sex)) %>%
  filter(grepl("1991", `Census Year`)) %>% 
  filter(grepl("services", `Detailed Industrial Group`)) %>% 
  select(`Detailed Industrial Group`, `Census Year`, value) %>% 
  arrange(desc(value))

Bringing It All Together

ggplot(industry_population %>% 
         filter(grepl("Garda|Defence", `Detailed Industrial Group`)) %>% 
         filter(grepl("Male|Female", Sex)) %>% 
         filter(!grepl("2016", `Census Year`)),
       aes(x = `Census Year`, y = value, fill = Sex)) + 
  facet_wrap(~`Detailed Industrial Group`, nrow = 1) +
  geom_bar(stat= "identity") +
  labs(title = "Population Working in Industrial Group", y = "Population") +
  dk_theme