Plotting with Matplotlib (Pandas)

Context: FIT1043_MOC · render the chart types in Python · matplotlib + pandas .plot · often after a groupby · lab: 30_Projects/FIT1043_Labs/Week4-Wrangling-Viz-Solution.pdf Problem it solves: turn a DataFrame column (or two) into the chart appropriate for its data type.

Quick Revision

• 🎯 Trigger: need a chart ➔ import matplotlib.pyplot as plt; call plt. or df.plot. matched to the data type.
• ⚡ Key Constraint: match plot to type — bar/pie for categorical, hist/boxplot/scatter for numeric.

🔧 Minimal Working Example

import matplotlib.pyplot as plt
df = pd.DataFrame({'Class': ['First','Second','Third'], 'Passengers': [194,177,450]})
plt.bar(df['Class'], df['Passengers'])   # categorical → bar chart
plt.show()

Expected output: a bar chart of passengers per class.

• Bar (categorical) ➔ plt.bar(df['Class'], df['Passengers']).
• Pie (categorical proportions) ➔ df.plot.pie(y='Average Age', labels=df['Class']).
• Line (trend) ➔ plt.plot(df['Y']).
• Scatter (cont × cont) ➔ plt.scatter(df['X'], df['Y']).
• Histogram (continuous) ➔ df.col_name.hist(bins=4).
• Boxplot (continuous) ➔ df.boxplot(column='col_name').

🔀 Variations

• Set the pie index ➔ build the DataFrame with index=['First','Second','Third'] so df.plot.pie(y='Average Age') labels sectors automatically.
• Control histogram detail ➔ raise/lower bins= (few = coarse shape, many = ragged).
• Encode a 3rd/4th variable on a scatter ➔ colour by a column and size the points: plt.scatter(df['HR'], df['SBP'], c=df['DBP'], s=40, cmap='hot') — c= colours, cmap= the palette, s= the marker size.

✍️ Practice

Practice 1: For numeric column price, draw a histogram with 20 bins and a boxplot to inspect its spread/outliers.

Reference solution

df['price'].hist(bins=20)          # distribution shape
df.boxplot(column='price')         # spread + outliers (IQR rule)

• Key move: histogram + boxplot are the numeric-continuous pair; both need a numeric column.

⚠️ Common Mistakes

• 💡 Wrong chart for the type ➔ a bar chart is for categorical counts; use a histogram for continuous data (it bins the values).
• 💡 plt.show() renders it ➔ in scripts the figure won’t display until plt.show() (notebooks may auto-render).