Level 4 · Are most orders small?

Bars compare, histograms spread

A bar chart compares categories. A histogram shows the distribution of one number: it cuts the range of values into intervals, called bins, and draws a bar for how many values fall into each. Neighboring bars touch, because the bins are next to each other on a number line.

Laura’s question, “is a typical order really $500?”, is about a distribution, so it gets a histogram of order values.

Drawing it

First one value per order (the starter does this), then ax.hist:

fig, ax = plt.subplots()
ax.hist(order_value, bins=range(0, 2100, 100), edgecolor="white")
ax.set_xlabel("Order value (USD)")
ax.set_ylabel("Orders")

bins takes either a number (bins=30 makes 30 equal bins from the smallest to the largest value) or the list of edges. With edges you control the range and the width. Here 21 edges make 20 bins of $100, from 0 to 2,000. Orders above $2,000, 207 of the 4,295 completed 2025 orders, fall outside the range and are not drawn. Between 20 and 50 bins usually works; with too few you lose the shape, with too many every bar is noise.

The numbers behind the bars

np.histogram takes the same arguments and returns what ax.hist draws:

counts, bin_edges = np.histogram(order_value, bins=range(0, 2100, 100))
counts[:3]        # [1180, 683, 451]
bin_edges[:3]     # [0, 100, 200]

There’s always one more edge than counts. Each bin includes its left edge and excludes its right edge, so an order of exactly $100 counts in the second bin. The last bin is the exception: it includes both ends.

Mean against median

This distribution is skewed to the right: lots of small orders and a long tail of big business baskets. In a skewed distribution the mean is pulled toward the tail, while the median stays where most orders are. Mark both with vertical lines and the gap is visible at once:

ax.axvline(order_value.mean(), color="red", label="Mean")             # $535.90
ax.axvline(order_value.median(), linestyle="--", label="Median")      # $257.96
ax.legend()

Half of 2025’s completed orders were under $258, yet the mean is more than double that. For “what does a typical customer spend”, the median is the honest answer; the mean is still the right number for revenue planning, because revenue is a total.

Common mistakes

• Plotting order_items lines instead of whole orders. Group to one value per order first.
• Leaving bins at the default of 10, which turns this shape into a few blocks.
• Using a bar chart with gaps between bars for a histogram, which suggests separate categories.
• Quoting only the mean of a skewed distribution. Give the median next to it.