Pokémon Data

Two days of real data analysis in Python. Day 1 answers questions about 800 Pokémon in a few lines of code. Day 2 flips it around: you invent a Pokémon and let the data argue with you about its stats.

Setup, if you have not done it: Python and VS Code Setup. Files: Datasets and Downloads.

Schedule

Project Day	Plan
1	Stats and Plots — explore `pokemon.csv` in a spreadsheet, then mean/median/min/max with pandas and a histogram, bar chart, and scatter plot with matplotlib
2	Design Your Own — use the Pokémon Designer tool to invent a Pokémon, compare it with box plots, scatter plots, and histograms, revise, save as CSV, sketch it

Project Day 1: Stats and Plots

Meet the Dataset

Download pokemon.csv.zip from Datasets, unzip it, and double-click pokemon.csv. It opens in Numbers or Excel.
Spend five minutes exploring. About how many rows? How many columns? What is your favorite Pokémon’s attack? Sort attack largest to smallest — who is on top? What is the difference between type1 and type2?

Install the Libraries

In the VS Code terminal, once:

python3 -m pip install pandas matplotlib

pandas works with data tables. matplotlib draws plots. Drag pokemon.csv into your python-class folder.

Basic Statistics

Create pokemon_stats.py in python-class:

import pandas as pd
import matplotlib.pyplot as plt

df = pd.read_csv("pokemon.csv")
print(df.head())

Run it. You should see the first five rows. df is a DataFrame, pandas’s word for a table. Then add:

print("Mean attack:   ", df["attack"].mean())
print("Median attack: ", df["attack"].median())
print("Min attack:    ", df["attack"].min())
print("Max attack:    ", df["attack"].max())
print(df["attack"].describe())

df["attack"] picks one column; .mean() asks for one number. .describe() prints the whole summary at once. Swap "attack" for "hp", "defense", or "speed" and run again.

Plots

A histogram shows the spread of one number column:

df["attack"].hist(bins=20)
plt.title("Attack Stat Distribution")
plt.xlabel("Attack")
plt.ylabel("Number of Pokémon")
plt.show()

A bar chart compares categories:

df["type1"].value_counts().plot(kind="bar")
plt.title("Pokémon Count by Primary Type")
plt.xlabel("Type")
plt.ylabel("Count")
plt.show()

A scatter plot compares two number columns, one dot per Pokémon:

df.plot.scatter(x="attack", y="defense")
plt.title("Attack vs. Defense")
plt.show()

Close each plot window before running again. Are high-attack Pokémon usually high-defense too?

Project Day 2: Design Your Own

Get the Designer

Download pokemon-designer.zip from Datasets and unzip it. You get a folder with designer.py and pokemon.csv.
File → Open Folder… and open pokemon-designer.
Optional, for hover tooltips: python3 -m pip install mplcursors. The tool runs without it.
Run designer.py. Type any name and type, reach the menu, and pick 5. Save and quit to confirm it works.

Design, Test, Revise

Before typing anything, decide on a personality (sneaky, tough, fast, fragile), a primary type and maybe a secondary, and a battle role — glass cannon, tank, speedster. The stats should match. A tank with speed = 200 makes no sense.

Run the tool again and enter your real design. Then use the menu:

Plot	What it shows	Best for
Box plot	Where your stat sits in the typical range for your type	“Is my attack normal for a fire type?”
Scatter plot	Your Pokémon as a star among all 800, on two stats	“Which real Pokémon is most like mine?”
Histogram	The distribution of one stat across every Pokémon	“Is my speed unusual anywhere?”

Try at least one of each. When a stat looks wrong, pick 4. Edit my Pokémon, fix it, and plot again. Loop until it fits — or until the outlier is on purpose.

Save and Sketch

Pick 5. Save and quit with a filename like your_pokemon_name.csv. Then, on paper, draw your Pokémon with its name, type(s), six stats, and a one-sentence description (“The Glitch Pokémon”).

Checkpoints

Day 1

I opened pokemon.csv in a spreadsheet and found one interesting fact.
pandas and matplotlib installed without errors.
I printed mean, median, min, and max for at least two columns.
I made a histogram and a bar chart, and changed at least one column name.

Day 2

I decided on a personality and role before entering numbers.
I tried a box plot, a scatter plot, and a histogram.
I revised at least one stat after looking at a plot.
I saved my Pokémon as a .csv and turned in a sketch.

Teacher Notes

The spreadsheet warmup on Day 1 matters: students need to see rows and columns before df["attack"] means anything.
pip failures are the main time sink. Have students run the install during the warmup so errors surface early.
Day 2 collects stats with a form in the original run; a CTLS submission works just as well.
The soccer dataset on the Datasets page is a ready-made extension for fast finishers.

Last updated on October 7, 2026