Lecture 12 · Pandas and Exploratory Data Analysis

Open the slides · Read the worked handout.

The 30-minute introduction covers DataFrames and Series, labels and positions, missing values, boolean selection, column arithmetic, summaries, and plots. It applies these operations to the saved NASA exoplanet table used in Project 4. The handout includes the two prediction answers.

Install Pandas¶

In the course terminal, activate your existing environment and install Pandas:

conda activate comp521
conda install -c conda-forge pandas

Review Conda’s proposed package plan before confirming. Use the existing course environment, rather than installing into base or mixing package managers. The Pandas installation guide provides the package reference.

Create your own notebook and load the data¶

Download the saved NASA CSV and its provenance record into your working folder. Start Jupyter there:

jupyter notebook

Create a blank Python notebook, choose your comp521 kernel, and save it in that folder. In your own code cell, load the table:

import pandas as pd

raw = pd.read_csv('project4-nasa-exoplanets.csv', comment='#')
raw.head()

The saved snapshot has 437 rows and 15 columns. Inspect raw.shape, raw.dtypes, and raw.isna().sum() to see what was loaded. Consult the NASA definitions for units and flags. No live archive query is needed.

Experiment with the table¶

Choose a question about the table. Try different columns, row selections, summaries, and plots using the lecture examples as references. Keep raw unchanged and save experiment selections under new variable names. Describe one result and one question to a neighbor. Counts describe the selected archive rows, not exoplanet occurrence rates.

You create the notebook yourself; there is no starter notebook to download. This is in-class exploration with no new submission. Project 4’s brief continues to define its required work.

Use Pandas fundamentals for further reference. Return to the Lecture Library.