Scatter plots

Scatter plots#

Documentation:
https://plotly.com/python/line-and-scatter/

Let’s start with the basic scatter plot from the earlier section of the tutorial. Here is what our data looks like:

import plotly.express as px
gap_data = px.data.gapminder()
just_2002 = gap_data.query('year == 2002')
just_2002.tail()

	country	continent	year	lifeExp	pop	gdpPercap	iso_alpha	iso_num
1654	Vietnam	Asia	2002	73.017	80908147	1764.456677	VNM	704
1666	West Bank and Gaza	Asia	2002	72.370	3389578	4515.487575	PSE	275
1678	Yemen, Rep.	Asia	2002	60.308	18701257	2234.820827	YEM	887
1690	Zambia	Africa	2002	39.193	10595811	1071.613938	ZMB	894
1702	Zimbabwe	Africa	2002	39.989	11926563	672.038623	ZWE	716

And here is the scatter plot:

scatter_2002 = px.scatter(
	data_frame=just_2002,
	x='gdpPercap',
	y='lifeExp',
)
scatter_2002

If we want to know about a particular data point, we can hover over it with the mouse. But we can also include additional formatting to make the information more visible. For example, instead of showing all data points in the same color, we can color the points by continent and give them different symbols as well.

Here are the continents in our data:

just_2002['continent'].unique()

array(['Asia', 'Europe', 'Africa', 'Americas', 'Oceania'], dtype=object)

In order to color the points by continent, we need to specify the color option in the scatter function. This option takes the name of a column in the data frame, and uses the values in that column to determine the color of each point.

scatter_2002 = px.scatter(
  data_frame=just_2002,
  x='gdpPercap',
  y='lifeExp',
  color='continent',
  symbol='continent',
)
scatter_2002

Let’s also simplify and improve the styling by doing the following:

use the simple_white template
add a title
change the axis and hover labels
add thin grey horizontal grid lines
remove the legend title
use the name of the country as the title in the hover box
add a dollar sign prefix to the GDP per capita axis

scatter_2002 = px.scatter(
  data_frame=just_2002,
  x='gdpPercap',
  y='lifeExp',
  color='continent',
  symbol='continent',
  template='simple_white',
  title='Relation between life expectancy & GDP per capita',
  hover_name='country',
  labels={
    'gdpPercap': 'GDP per capita (USD)',
    'lifeExp': 'Life expectancy (years)',
    'continent': 'Continent',
    'country': 'Country',
  },
)
scatter_2002.update_yaxes(
  showgrid=True,
	# set the width in pixels for the gridlines
	gridwidth=1,
	# set the grid color
	gridcolor='lightgrey',
)
scatter_2002.update_xaxes(
  tickprefix='$',
)
scatter_2002.update_layout(
  legend_title=None,
)
scatter_2002

Let’s do one last thing. Let’s scale the size of the points by the population of each country.

How does this affect the readability of the chart?

scatter_2002 = px.scatter(
  data_frame=just_2002,
  x='gdpPercap',
  y='lifeExp',
  color='continent',
  symbol='continent',
  template='simple_white',
  title='Relation between life expectancy & GDP per capita',
  hover_name='country',
  labels={
    'gdpPercap': 'GDP per capita (USD)',
    'lifeExp': 'Life expectancy (years)',
    'continent': 'Continent',
    'country': 'Country',
  },
  size='pop',
)
scatter_2002.update_yaxes(
  showgrid=True,
	# set the width in pixels for the gridlines
	gridwidth=1,
	# set the grid color
	gridcolor='lightgrey',
)
scatter_2002.update_xaxes(
  tickprefix='$',
)
scatter_2002.update_layout(
  legend_title=None,
)
scatter_2002