Lab 1: Computing Setup

Overview

Today we’re going to install (or double check installations for) the computing tools we will use in this course and practice the lab report submission workflow. If you are an experienced programmer, you are free to follow your own workflow and/or use your own toolkit for implementing labs - but there is a lower probability that I can help you troubleshoot any issues you run into. Today’s lab is not graded.

Today we will:

1. Install VS Code, Mamba, Python, and git.
2. Setup our GitHub account and create a repository for today’s lab.
3. Run through the setup steps typical of future labs, run some code, export our notebook, and submit a pdf on Canvas.

Lab Workflow

Installation and Setup

1. Download VS Code for your machine if you don’t already have it.
2. Install Mamba for your system if you don’t already have it. We will use Python in individual Mamba environments as opposed to a system-wide distribution.
   • Please review key concepts for environment management with Mamba.
3. Install the Jupyter & Python VS Code extensions, along with their extension packs.
4. Download git for your machine if you don’t already have it.
5. Setup a GitHub account if you don’t already have one. I highly recommend following the instructions for Connecting to GitHub with SSH.
6. Create a new Public repository called ‘ENGS 199.20 Lab 1’. Add a README, .gitignore, and License. Learn about these files if you are unfamiliar with them.
7. Clone your repository to your local machine. I recommend having a directory for this class and to clone the repository inside that directory.
8. Download this file as a Jupyter notebook (see the upper righthand corner of the page) and add it to your project directory. You can use whatever git workflow you’d like to do version control, but I will provide instructions for using the command line.
   • Run git remote -v to check that your local repository is tracking the repository on GitHub.
   • Run git status from this lab directory’s root to see that the notebook is untracked.
   • Run git add . to track the notebook.
   • Run git commit -m [message] with your own helpful message.
   • Use this simple workflow when you make changes to files. If you don’t want to track a file, you can add the specific file’s location (or use wildcards and file prefixes or suffixes) to avoid tracking.
9. For this lab, we are going to create a small environment that allows us to plot using the seaborn package. After following these next instructions, we will be ready to run our notebook.

Sometimes the steps above can cause trouble! Please loop me into these struggles so we can troubleshoot together and save you time. It’s also a good opportunity to get to know the very nice and smart folks at IT who are experts with these setup problems!

Create a file called env-lab1.yml in your lab1 directory. Add the following contents and save:

name: lab1
channels:
  - conda-forge
dependencies:
  - python
  - nbconvert
  - ipykernel
  - seaborn
  - jupyter

Run mamba env create -f env-lab1.yml. Run mamba activate lab1. Set up the kernel for the environment with python -m ipykernel install --user --name lab1 --display-name lab1.

Run Code

Below, I reproduce a plot from the seaborn Gallery. Make sure the lab1 kernel is set up in the notebook and then run the code block.

import seaborn as sns
sns.set_theme(style="ticks", palette="pastel")

# Load the example tips dataset
tips = sns.load_dataset("tips")

# Draw a nested boxplot to show bills by day and time
sns.boxplot(x="day", y="total_bill",
            hue="smoker", palette=["m", "g"],
            data=tips)
sns.despine(offset=10, trim=True)

Attractive and clear figures are very important in decision analysis. Run the cell below, which makes the figure higher resolution and makes the font more legible. Some changes reflect my personal style preferences.

import matplotlib.pyplot as plt

# Set up a figure with higher resolution
fig, ax = plt.subplots(dpi=300)

# Draw a nested boxplot to show bills by day and time
sns.boxplot(x="day",
            y="total_bill",
            hue="smoker",
            showfliers=False,
            palette=["m", "g"],
            showmeans=True,
            meanprops={'marker': 'D',
                       'markerfacecolor': 'firebrick',
                       'markeredgecolor': 'firebrick'},
            data=tips,
            ax=ax)

# Rename legend title and increase font size
sns.move_legend(ax,
                loc='best',
                title='Smoker at Table?',
                alignment='left',
                fontsize=12,
                title_fontsize=14)

# Make tick labels more legible
ax.tick_params(labelsize=12)

# Rename axis labels
ax.set_ylabel("Total Bill ($)", size=14)
ax.set_xlabel("Day", size=14)
plt.show()

The distribution of restaurant bills in dollars reported by a waiter over a period of several months, shown in boxplots. The grouping examines how total bill size is distributed across day of the week (x-axis) and whether the party included a smoker (hue). Red diamonds denote the mean restaurant bill in each group.

Following the link in the seaborn example, we can learn more about the tips data here.

To complete this lab report, add a code cell where you create a new figure that explores the relationship between variables in this dataset. It does not have to use the variables we used so far, nor do you have to use a boxplot. Make sure your tick and axis labels are at least size 12 and that the figure resolution is at least 300 dots per inch. Make sure your figure includes a caption, which must state the association considered in the exploratory visualization. In the cell after your plot, add a markdown cell where you:

1. Write out a caption;
2. Describe the plotted result;
3. Suggest further analysis that could help uncover the relationship between these and/or other variables.

See a text example below for the revised figure from the seaborn example (note the caption underneath the revised figure on the website).

Analysis: On average, the waiter’s bills were similar regardless of day of the week or whether a smoker was present in the party. On Saturdays and Sundays, bills tended to be larger and more variable when smokers are in a party.

Future work: Lacking information on the relative frequency of bills across days of the week and smokers present in a party, it is difficult to draw strong conclusions about the associations of bill size and the considered variables. To better understand these associations, it would be helpful to show histograms in a second panel under the present plot. In addition, it’s not clear why presence of smokers in a dining party would have any association with the bill size. Other variables in the dataset, such as time of day and size of the party are likely more relevant to consider.

Note: If you do not have experience with Python, I recommend trying to generate a new plot – the seaborn package does not really require Python knowledge to generate a new plot. However, today’s lab is not graded so you don’t have to generate a new plot, write a caption, etc., I think it’s a good opportunity to practice the lab submission workflow and ask questions now before I start grading labs.

Submit Code and Report

Make sure your GitHub repository is up to date.

Export your .ipynb as a .pdf and submit on Canvas. In your submission, share a link to your repository. Please replace “ENGS 199.20” in the top cell with your name and change the filename to include your name.

Printing Code to PDF

Because we included nbconvert in our computational environment, we should be able to export the .ipynb to .pdf using VS Code and get a nicely formatted file.