Creating Flourish graphs in R and Python

Flourish primarily depends on two main functions in R and Python.

• Python users: Flourish() and a bind_..._data() function.
• R users: flourish() and a bind_..._data() function.

The ... depends on the chart type. A list of chart types can be found flourish_api_documentation['chart_type'] in R or reference['chart_type'] in Python.

In Python you can wrap the function in help() to see the docstrings and also shift tab to view the docstrings dynamically while coding. In R, you can find a list of Flourish functions by running flourishcharts::, typing ?flourishcharts in your console, or heading over to the references page in this site.

The following chunks show some example charts using Gapminder data, with accompanying notes.

The Flourish Charts package provides a data frame (flourish_api_documentation in R or reference in Python) with metadata about each graph (template ID, version, a simplified chart name for end users, and a URL to Flourish's API documentation).

from flourishcharts import Flourish
from flourishcharts.datasets import gapminder
from flourishcharts.flourish_docs import reference
import pandas as pd

library(flourishcharts)
library(tidyr)
library(dplyr)

reference[['template_name', 'chart_type', 'url']].head()

head(flourish_api_documentation)

Flourish Charts requires the following functions to plot Flourish graphs:

• Mandatory arguments in flourish() in R or Flourish() in Python:
  • chart_type: A string that must match a value in flourish_api_documentation["chart_type"]. This tells Flourish the chart template you want to use to visualize your data.
  • The API key. This can be set in two ways:
    • As per the setup steps in Flourish API key which will point to the default value.
    • or by passing a string to api_key.
• bind_..._data(): A data bindings function that allows users to upload data to the graph and tell Flourish which column, or data binding, should go where on the chart. For more information on the information required, see the Flourish API documentation for each graph type⁠(opens in a new tab or window).
• Add additional details by running set_..._details() — for example, set_scatter_details(). This is where you can change the chart color palette, axes labels, and so on.

Every function can be piped or chained into another. For example, R users can use the base pipe |> to chain together the main flourish() function and a data bindings function, like bind_scatter_data : flourish() |> bind_scatter_data(). Python users are able to chain the functions together by appending each function prefixed with a period: Flourish().bind_scatter_data().

scatterplot = Flourish(
    chart_type = "scatter"
).bind_scatter_data(
    data = gapminder,
    x = "gdpPercap",
    y = "lifeExp",
    slider = "year",
    metadata = ["country", "year"],
    size = "pop",
    color = "continent"
)
scatterplot

scatterplot <- flourish(
  chart_type = "scatter"
) |>
bind_scatter_data(
  data = gapminder,
  x = "gdpPercap",
  y = "lifeExp",
  slider = "year",
  size = "pop",
  color = "continent",
  metadata = c("country", "year")
)
scatterplot

Please note that some Flourish graphs require wide datasets. We convert the gapminder dataframe from long to wide below.

# Define the subset of countries
subset_countries = [
    "Australia", "New Zealand", "United States",
    "Rwanda", "Sierra Leone", "Indonesia", "Brazil"
]
# Filter the DataFrame and select relevant columns
subset_countries_bcr_df = (
    gapminder[gapminder['country'].isin(subset_countries)]
    [["year", "lifeExp", "country", "continent"]]
)
# Pivot the DataFrame to a wide format
countries_wide_bcr = subset_countries_bcr_df.pivot(
    index=["country", "continent"],
    columns="year",
    values="lifeExp"
)
# Convert the pivot table to a DataFrame
countries_wide_bcr_df = countries_wide_bcr.reset_index()

bcr = Flourish(
    "bar_race"
).bind_bar_chart_race_data(
    data = countries_wide_bcr_df,
    label = "country",
    values = ["1952", "1957",
    "1962", "1967", "1972", "1977",
    "1982", "1987", "1992", "1997",
    "2002", "2007"],
    category = "continent"
).set_bar_chart_race_details(
    chart_layout_title = "Life expectancy from the 1950s to 2007",
    chart_layout_subtitle = "Selected countries include Australia, New Zealand, the US, Rwanda, Indonesia, Sierra Leone, and Brazil.",
    totaliser = False
)
bcr

bcr_data <- gapminder |>
  filter(country %in% c(
    "Australia",
    "New Zealand",
    "United States",
    "Rwanda",
    "Sierra Leone",
    "Indonesia",
    "Brazil"
  )) |>
  select(c("country", "continent", "year", "lifeExp")) |>
  pivot_wider(
    id_cols = c("country", "continent"),
    names_from = "year",
    values_from = "lifeExp"
)

bcr <- flourish("bar_race") |>
  bind_bar_chart_race_data(
    data = bcr_data,
    label = "country",
    values = colnames(bcr_data[, c(3:14)]),
    category = "continent"
  ) |>
  set_bar_chart_race_details(
    chart_layout_title = "Life expectancy from the 1950s to 2007",
    chart_layout_subtitle = "Selected countries include Australia, New Zealand, the US, Rwanda, Indonesia, Sierra Leone, and Brazil.",
    totaliser = FALSE
  )
bcr

from io import StringIO

d = '''before,after,seat_flow
Radicals,Radicals,27
Moderates,Moderates,18
Green,Green,4
Progressives,Progressives,15
Radicals,Moderates,28
Progressives,Radicals,12
Libertarians,Libertarians,8
Moderates,Radicals,10
Progressives,Moderates,6
Radicals,Independents,5
Independents,Independents,4
Progressives,Independents,3
Progressives,Libertarians,7
Independents,Moderates,2
Loonies,Green,2
Independents,Radicals,1
Loonies,Libertarians,10
Independents,Independents,1'''


df = pd.read_csv(StringIO(d), sep=',')

alluvial_graph = Flourish(
    chart_type = "sankey"
).bind_sankey_data(
    data = df,
    source = "before",
    target = "after",
    value = "seat_flow"
)
alluvial_graph

alluvial_data <- tibble::tribble(
    ~before,~after,~seat_flow,
    "Radicals","Radicals" , 27,
    "Moderates","Moderates",1,
    "Green","Green",4,
    "Progressives","Progressives",15,
    "Radicals","Moderates",28,
    "Progressives","Radicals",12,
    "Libertarians","Libertarians",8,
    "Moderates","Radicals",10,
    "Progressives","Moderates",6,
    "Radicals","Independents",5,
    "Independents","Independents",4,
    "Progressives","Independents",3,
    "Progressives","Libertarians",7,
    "Independents","Moderates",2,
    "Loonies","Green",2,
    "Independents","Radicals",1,
    "Loonies","Libertarians",10,
    "Independents","Independents",1
)

alluvial_graph <- flourish(chart_type = "sankey") |>
  bind_sankey_data(data = alluvial_data,
  source = "before",
  target = "after",
  value = "seat_flow"
)
alluvial_graph