Exercise 07

Making figures with Python

Published

April 1, 2025

Modified

April 1, 2025

Work in progress

This page is under construction.

It may change before the assignment is released to class.

Dates

We’ll work on this exercise in-class on Thursday, April 10, 2025.

The write-up is due on Thursday, April 17, 2025.

Goals

Create some simple figures in Python using the Pandas library.
Gain an appreciation of the costs and benefits of scripting the generation of figures.

Assignment

Set-up

For reasons not especially worth explaning here, we have to use R to configure Python for making figures using posit.cloud.

Code

library(reticulate)
py_require(c("pandas", "matplotlib", "numpy"))

Python calls groups of functions libraries. These are analogous to packages in R.

Code

import numpy as np
import matplotlib.pyplot as plt
import pandas as pd

Note

In Python, we use the import command.

We create ‘nicknames’ for the packages so that we can refer to them using an easy-to-type shorthand. The nicknames are the short names: … import pandas as pd, means import the ‘pandas’ library and give it the shortname of ‘pd’.

NFSG data

Gather

We’ll make plots of some of NSFG data we discussed in class on 2025-04-01.

Note

The data file has been saved under csv/

The pandas library (shortname in our code pd) handles the creation and manipulation of data frames. That includes importing comma-separated value (CSV) files.

Code

nsfg = pd.read_csv('../include/csv/NSFG_2022_2023_FemPregPUFData.csv')

We confirm that this worked by checking the data types in nsfg:

Code

nsfg.dtypes

CaseID           int64
PREGORDR         int64
FTFMODE          int64
BORNALIV       float64
RECNT5YRPRG    float64
                ...   
CMJAN3YR         int64
CMJAN4YR         int64
CMJAN5YR         int64
YEAR             int64
QUARTER          int64
Length: 111, dtype: object

This is similar to running the str() function on an R data frame.

Python methods

Objects in Python have specialized functions that can be used with them using a simple ‘dot’ syntax. So nsfg.dtypes means ‘run the data types function on the nsfg data frame.’ These specialized functions are called ‘methods.’

Code

nsfg.shape

(8247, 111)

The shape method is similar to the dim() function in R. What do the two numbers mean?

Plot

Since we used the pandas library to import our data frame, we can use one of the built-in methods that apply to data frames to plot a histogram. Here, we create a histogram by calling the hist() method on the nsfg data frame and by specifying the column AGER, the age of the responding participant.

Code

nsfg.hist(column = "AGER")

array([[<Axes: title={'center': 'AGER'}>]], dtype=object)

Code

plt.show()

Now, let’s customize the plot by changing some parameters in the hist() method. Add change the number of bins to some larger number like 20, 25, or 30 (the default is 10), by changing LARGE_NUMBER to a number.

Code

LARGE_NUMBER = 35
nsfg.hist(column = "AGER", bins = LARGE_NUMBER)

array([[<Axes: title={'center': 'AGER'}>]], dtype=object)

Code

plt.show()

Now, try a smaller value, less than 10. Change the code below to try this.

Code

SMALLER_NUMBER = 5
nsfg.hist(column = "AGER", bins = SMALLER_NUMBER)

array([[<Axes: title={'center': 'AGER'}>]], dtype=object)

Code

plt.show()

What do you notice?

Let’s look at the histograms by RELIGION, like we did in class on 2025-04-01.

Code

nsfg.hist(column = 'AGER', by = 'RELIGION')

array([[<Axes: title={'center': '1'}>, <Axes: title={'center': '2'}>],
       [<Axes: title={'center': '3'}>, <Axes: title={'center': '4'}>]],
      dtype=object)

Code

plt.show()

Modify the code below to create a set of histograms by some other variable that you choose (change VARIABLE_YOU_CHOOSE in the code below.) Make sure to look at the codebook to make sure that the variable you choose makes sense.

Warning

Make sure to put the variable you choose in quotations.

Code

VARIABLE_YOU_CHOOSE = 'BABYSEX'
nsfg.hist(column = 'AGER', by = VARIABLE_YOU_CHOOSE)

array([[<Axes: title={'center': '1.0'}>, <Axes: title={'center': '2.0'}>],
       [<Axes: title={'center': '8.0'}>, <Axes: title={'center': '9.0'}>]],
      dtype=object)

Code

plt.show()

Finally, experiment with changing some of the default parameters like grid (values can be True or False), xrot (rotation of x axis labels) or yrot (rotation of y axis labels).

Code

MY_XROT = 0
nsfg.hist(column = 'AGER', by = 'RELIGION', xrot = MY_XROT)

array([[<Axes: title={'center': '1'}>, <Axes: title={'center': '2'}>],
       [<Axes: title={'center': '3'}>, <Axes: title={'center': '4'}>]],
      dtype=object)

Code

plt.show()

Plot other

Submit

The code you wrote in following the steps above.
The results of running your code.
Comments about what you observed.

--- title: "Exercise 07" subtitle: "Making figures with Python" format: html: code-fold: true docx: default --- ::: {.callout-warning} # Work in progress This page is under construction. It may change before the assignment is released to class. ::: ## Dates {-} We'll work on this exercise in-class on Thursday, April 10, 2025. The write-up is [due]{.orange_due} on Thursday, April 17, 2025. ## Goals {-} 1. Create some simple figures in Python using the Pandas library. 2. Gain an appreciation of the costs and benefits of scripting the generation of figures. ## Assignment {-} ## Set-up For reasons not especially worth explaning here, we have to use R to configure Python for making figures using posit.cloud. ```{r} #| label: set-up-python-env library(reticulate) py_require(c("pandas", "matplotlib", "numpy")) ``` Python calls groups of functions *libraries*. These are analogous to *packages* in R. ```{python} import numpy as np import matplotlib.pyplot as plt import pandas as pd ``` ::: {.callout-note} In Python, we use the `import` command. We create 'nicknames' for the packages so that we can refer to them using an easy-to-type shorthand. The nicknames are the short names: ... `import pandas as pd`, means import the 'pandas' library and give it the shortname of 'pd'. ::: ## NFSG data ### Gather We'll make plots of some of NSFG data we discussed in class on [2025-04-01](../slides/wk11-04-01-more-ggplot.qmd). ::: {.callout-note} The data file has been saved under `csv/` ::: The `pandas` library (shortname in our code `pd`) handles the creation and manipulation of data frames. That includes importing comma-separated value (CSV) files. ```{python} #| label: import-survey-data-csv #| echo: true nsfg = pd.read_csv('../include/csv/NSFG_2022_2023_FemPregPUFData.csv') ``` We confirm that this worked by checking the data types in `nsfg`: ```{python} #| label: check-data-types #| echo: true nsfg.dtypes ``` This is similar to running the `str()` function on an R data frame. ::: {.callout-important} ## Python methods Objects in Python have specialized functions that can be used with them using a simple 'dot' syntax. So `nsfg.dtypes` means 'run the data types function on the nsfg data frame.' These specialized functions are called 'methods.' ::: ```{python} #| label: nsfg-shape #| echo: true nsfg.shape ``` The `shape` method is similar to the `dim()` function in R. What do the two numbers mean? ### Plot Since we used the `pandas` library to import our data frame, we can use one of the built-in methods that apply to data frames to plot a histogram. Here, we create a histogram by calling the `hist()` method on the `nsfg` data frame and by specifying the column `AGER`, the age of the responding participant. ```{python} #| label: nsfg-hist #| echo: true nsfg.hist(column = "AGER") plt.show() ``` Now, let's customize the plot by changing some parameters in the `hist()` method. Add change the number of bins to some larger number like 20, 25, or 30 (the default is 10), by changing `LARGE_NUMBER` to a number. ```{python} #| label: change-bins #| echo: true LARGE_NUMBER = 35 nsfg.hist(column = "AGER", bins = LARGE_NUMBER) plt.show() ``` Now, try a smaller value, less than 10. Change the code below to try this. ```{python} #| label: change-bins-smaller #| echo: true SMALLER_NUMBER = 5 nsfg.hist(column = "AGER", bins = SMALLER_NUMBER) plt.show() ``` What do you notice? Let's look at the histograms by `RELIGION`, like we did in class on [2025-04-01](https://psu-psychology.github.io/psych-490-data-viz-2025-spring/slides/wk11-2025-04-01-more-ggplot.html). ```{python} #| label: hist-ager-by-religion #| echo: true nsfg.hist(column = 'AGER', by = 'RELIGION') plt.show() ``` Modify the code below to create a set of histograms by some *other* variable that you choose (change `VARIABLE_YOU_CHOOSE` in the code below.) Make sure to look at the [codebook](https://psu-psychology.github.io/psych-490-data-viz-2025-spring/include/pdf/2022-2023-NSFG-FemRespPUFCodebook.pdf) to make sure that the variable you choose makes sense. ::: {.callout-warning} Make sure to put the variable you choose in quotations. ::: ```{python} #| label: hist-students-choice #| echo: true VARIABLE_YOU_CHOOSE = 'BABYSEX' nsfg.hist(column = 'AGER', by = VARIABLE_YOU_CHOOSE) plt.show() ``` Finally, experiment with changing some of the default parameters like `grid` (values can be `True` or `False`), `xrot` (rotation of x axis labels) or `yrot` (rotation of y axis labels). ```{python} #| label: hist-students-choice-2 #| echo: true MY_XROT = 0 nsfg.hist(column = 'AGER', by = 'RELIGION', xrot = MY_XROT) plt.show() ``` ### Plot other ## Submit {-} 1. The code you wrote in following the steps above. 2. The results of running your code. 3. Comments about what you observed.