P1D6: Open programming time for Python

First principles weighted mean

1. Let’s talk about lambda functions
2. Why we want to use them with df.assign()
3. Working through a groupby

Plotnine’s port of ggplot2

What are the key differences you have noted between ggplot2 and plotnine?

• Quoting table columns
• .save() method versus the ggsave() function.

Questions about Altair

• alt.Chart() and the introduction
• alt.X()
• alt.Y()
• alt.Scale()
• alt.Axis()
• alt.Size()
• alt.Column()
• alt.layer()

  alt.Chart() methods

alt.Chart() and the introduction

• .encode()
• mark_<type>()
• .properties()
• .facet()
• .configure_header()` and the other configures

Note that Altair defaults to placing commas at the thousands. We can fix that with alt.Axis(format='.0f')

You can only use one data object

You can’t facet with multiple data sets. So, we need to get our continent data into our country DataFrame.

• I create a new continent_color column.
• Make sure the columns are identically named.
• In cont_facet I create a country column with 'Continent' repeated. This will help in the detail argument of .mark_line().
• In cont_facet then create a "Continent" value in the continent_color column. This will allow me to mark it on the legend as the continent average in black.
• Then we leverage the .append() method from Pandas.

count_facet = (gapminder
    .filter(['country', 'continent', 'year',
    'pop', 'gdpPercap', 'pop100k'], axis=1)
    .assign(continent_color = lambda x: x.continent)
)
cont_facet = (gm_con
    .filter(['continent', 'year', 'sum_pop', 
        'gdpPercap', 'pop100k'], axis=1)
    .rename({'sum_pop':'pop'},
        axis='columns')
    .assign(
        country = "Continent",
        continent_color = "Continent" 
         )
)

dat_facet = count_facet.append(cont_facet)