Level 1 · Black Friday in one line

Start from the question, then the data

A chart answers one question. Megan’s is “how did orders move day by day through the holidays, and how big was the peak?” Time on the x-axis, a count on the y-axis, points that belong in order: that is a line chart.

Before drawing anything, build the numbers. One row per day:

season = orders[(orders["order_date"] >= "2025-11-01") & (orders["order_date"] < "2026-01-01")]
daily = season.set_index("order_date").resample("D").size().rename("n_orders")
len(daily)       # 61 days
daily.max()      # 67, on 2025-12-01 (Cyber Monday)
daily.median()   # 22

The chart is only as good as this Series, which is why the check looks at it.

Figure and Axes

matplotlib draws on two objects. The Figure is the whole image, the frame. The Axes is the plotting area inside it, with its x and y axis and the data. plt.subplots() makes one of each:

import matplotlib.pyplot as plt

fig, ax = plt.subplots(figsize=(9, 4))   # width, height in inches
ax.plot(daily.index, daily.values)       # x first, then y
ax.set_title("Orders per day, Nov to Dec 2025")
ax.set_xlabel("Day")
ax.set_ylabel("Orders")

x and y must have the same length; here both come from the same Series, so they always do.

You’ll also see a shorter style that talks to “the current chart”: plt.plot(...), plt.title(...), plt.xlabel(...). It’s fine for a quick look. The fig, ax style is the one to learn, because it’s clearer and you need it as soon as a figure holds more than one chart. The names map one to one: plt.title becomes ax.set_title, plt.xlabel becomes ax.set_xlabel, plt.ylim becomes ax.set_ylim.

pandas has a shortcut too: daily.plot() draws the same line and returns the ax, so you can keep going with ax.set_title(...).

Reference lines and markers

A line alone shows the shape. A reference line gives it a scale:

ax.plot(daily.index, daily.values, marker="o", markersize=3)
ax.axhline(daily.mean(), color="gray", linestyle="--", label="Average")
ax.legend()

axhline draws a horizontal line across the chart (axvline a vertical one, handy for marking a launch date). marker="o" puts a dot on each day; linestyle takes "-", "--" or ":". ax.legend() lists every line that was given a label=. Dates along the x-axis often run into each other; fig.autofmt_xdate() tilts them so they fit.

Here the average is about 24 orders a day. The Black Friday weekend (November 28 to 30) brought 56, 51 and 58, and Cyber Monday, December 1, brought 67: almost three times a normal day. The quietest days were Christmas Eve and Christmas Day, with 11 and 9.

A title that says the finding

“Orders per day” describes the chart. “Cyber Monday was the busiest day of the season” tells the reader what to see. A good chart can be understood without you standing next to it, and the title is where that starts.

Common mistakes

• Calling ax.title("...") instead of ax.set_title("..."). On an Axes the setters start with set_; ax.title is a text object, and calling it raises TypeError: 'Text' object is not callable.
• Calling ax.legend() when no line has a label=. The legend comes out empty with a warning.
• Removing the import matplotlib.pyplot as plt line. Here that line is what loads the chart library, and pandas’ .plot() needs it too.