Plotting Data Lab

Learning Objectives

Understand the components of a point in a graph, an $x$ value, and a $y$ value
Understand how to plot a point on a graph, from a point's $x$ and $y$ value
Get a sense of how to use a graphing library, like Plotly, to answer questions about our data

Working again with our travel data

Let's again get our travel data from our excel spreadsheet. If we do not already have pandas and xldr for retrieving data from excel, we should install it now.

!pip install pandas
!pip install xlrd

import pandas
file_name = './cities.xlsx'
travel_df = pandas.read_excel(file_name)
cities = travel_df.to_dict('records')

Press shift + enter to run the code above.

cities

[{'Area': 4758,
  'City': 'Buenos Aires',
  'Country': 'Argentina',
  'Population': 2891000},
 {'Area': 2731571,
  'City': 'Toronto',
  'Country': 'Canada',
  'Population': 2800000},
 {'Area': 3194,
  'City': 'Pyeongchang',
  'Country': 'South Korea',
  'Population': 2581000},
 {'Area': 200, 'City': 'Marakesh', 'Country': 'Morocco', 'Population': 928850},
 {'Area': 491,
  'City': 'Albuquerque',
  'Country': 'New Mexico',
  'Population': 559277},
 {'Area': 3750,
  'City': 'Los Cabos',
  'Country': 'Mexico',
  'Population': 287651},
 {'Area': 68, 'City': 'Greenville', 'Country': 'USA', 'Population': 84554},
 {'Area': 8300,
  'City': 'Archipelago Sea',
  'Country': 'Finland',
  'Population': 60000},
 {'Area': 33,
  'City': 'Walla Walla Valley',
  'Country': 'USA',
  'Population': 32237},
 {'Area': 27, 'City': 'Salina Island', 'Country': 'Italy', 'Population': 4000},
 {'Area': 59, 'City': 'Solta', 'Country': 'Croatia', 'Population': 1700},
 {'Area': 672,
  'City': 'Iguazu Falls',
  'Country': 'Argentina',
  'Population': 0}]

Plotting our first graph

As we can see, in our list of cities, each city has a population number. Our first task will be to display the populations of our first three cities in a bar chart.

First we load the plotly library into our notebook, and we initialize this offline mode.

import plotly

plotly.offline.init_notebook_mode(connected=True)
# use offline mode to avoid initial registration

Now the next step is to build a trace. As we know our trace is a dictionary with a key of x and a key of y. We have set up a trace to look like the following: trace_first_three = {'x': x_values, 'y': y_values}.

First define x_values so that it is a list of names of the first three cities. Use what we learned about accessing information from lists and dictionaries to assign x_values equal to the first three cities.

x_values = [cities[0]['City'], cities[1]['City'], cities[2]['City']]

Now use list and dictionary accessors to set y_values equal to the first three populations.

y_values = [cities[0]['Population'], cities[1]['Population'], cities[2]['Population']]

Now let's plot our data.

trace_first_three_pops = {'x': x_values, 'y': y_values}


plotly.offline.iplot([trace_first_three_pops])

Modifying our first trace

Note that by default, plotly sets the type of trace as a line trace. In the next example, let's make our trace a bar trace by setting the 'type' key equal to 'bar'. We can continue to use our lists of x_values and y_values that we defined above and used in our previous trace. To make our new trace more informative, we can assign labels to our data when we plot it. Normally, when we see a bar graph, there are labels along the x-axis for specific values. Understanding that we are plotting data about different cities, our labels would sensibly be a list of corresponding city names.

We can designate these corresponding city names in our trace dictionary by assigning a list of strings to the text key:

example_trace = {'type': 'bar', 'x': x_values, 'y': y_values, 'text': ["label_1", "label_2", "label_3"]}

Assign the variable text_values equal to a list of names for the first three cities. Then we pass this information to our trace dictionary and assign it as the value for its text key.

text_values = []
bar_trace_first_three_pops = {'type': 'scatter', 'text': text_values}

bar_trace_first_three_pops['type'] # 'bar'

plotly.offline.iplot([bar_trace_first_three_pops])

Adding a second trace to plot side by side

Ok, now let's plot two different traces side by side. First, create another trace called bar_trace_first_three_areas that is like our bar_trace_first_three_pops except the values are a list of areas. We will display this side by side along our bar_trace_first_three_pops in the plot below.

bar_trace_first_three_areas = {'type': 'scatter', 'x': [], 'y': [], 'text': []}
bar_trace_first_three_pops = {'type': 'scatter', 'x': [], 'y': [], 'text': []}

plotly.offline.iplot([bar_trace_first_three_pops, bar_trace_first_three_areas])

Summary

In this section, we saw how we use data visualizations to better understand the data. We do the following. Import plotly:

import plotly

plotly.offline.init_notebook_mode(connected=True)

Then we define a trace, which is a Python dictionary.

trace = {'x': [], 'y': [], 'text': [], 'type': 'bar'}

Finally, we display our trace with a call to the following method:

plotly.offline.iplot([trace])

Easy peasy, quick and easy!