PY-L4-34 · Matplotlib — Line Charts & Time Series

Learning Goals

3 min

Draw a basic line chart with ax.plot(x, y).
Plot multiple series on the same axes with a legend.
Format date axes with matplotlib.dates locators.
Overlay a 7-day rolling average to smooth a noisy series.

Warm-Up · A 1-Line Plot

5 min

import matplotlib.pyplot as plt
days = list(range(1, 11))
sales = [3, 5, 4, 6, 8, 7, 9, 11, 10, 12]
plt.plot(days, sales)
plt.show()

Today's big idea

Lines are for ordered data — usually time. The x-axis tells your story; pick a sensible range, format dates legibly, and only show as many ticks as the eye can read.

New Concept · Lines, Dates, Smoothing

14 min

Line style + markers

ax.plot(x, y, color="#4a90e2", linewidth=2, marker="o", label="actual")

Multiple series

ax.plot(x, y1, label="Apr")
ax.plot(x, y2, label="May")
ax.legend()

Date axis

import matplotlib.dates as mdates
import pandas as pd

df = pd.read_csv("daily.csv", parse_dates=["date"])
fig, ax = plt.subplots(figsize=(10, 5))
ax.plot(df["date"], df["sales"])

# major tick every month, format as "May 2026"
ax.xaxis.set_major_locator(mdates.MonthLocator())
ax.xaxis.set_major_formatter(mdates.DateFormatter("%b %Y"))
fig.autofmt_xdate()

autofmt_xdate rotates labels so they don't overlap.

Rolling average

df["rolling7"] = df["sales"].rolling(7).mean()
ax.plot(df["date"], df["sales"], alpha=0.4, label="daily")
ax.plot(df["date"], df["rolling7"], linewidth=2, label="7-day avg")

The translucent daily line + the bold rolling line shows both signal and noise — a great default for any noisy daily metric.

Annotate a key moment

peak = df.loc[df["sales"].idxmax()]
ax.annotate(f"peak: {int(peak['sales'])}",
            xy=(peak["date"], peak["sales"]),
            xytext=(peak["date"], peak["sales"] + 30),
            arrowprops={"arrowstyle": "->"})

Worked Example · Daily Revenue with Trend Line

12 min

# daily_revenue.py
import pandas as pd
import matplotlib.pyplot as plt
import matplotlib.dates as mdates

df = pd.read_csv("clean.csv", parse_dates=["date"])
df["total"] = df["quantity"] * df["price"]

daily = (df.groupby(df["date"].dt.date)["total"].sum()
           .rename("daily").to_frame())
daily.index = pd.to_datetime(daily.index)
daily["roll7"] = daily["daily"].rolling(7).mean()

fig, ax = plt.subplots(figsize=(10, 5))
ax.plot(daily.index, daily["daily"],
        color="#4a90e2", alpha=0.35, label="daily")
ax.plot(daily.index, daily["roll7"],
        color="#1f5fa8", linewidth=2.5, label="7-day avg")

ax.set_title("Daily revenue · 7-day rolling")
ax.set_ylabel("RM")
ax.xaxis.set_major_locator(mdates.WeekdayLocator(byweekday=mdates.MO))
ax.xaxis.set_major_formatter(mdates.DateFormatter("%d %b"))
ax.grid(axis="y", linestyle="--", alpha=0.4)
ax.legend()

fig.autofmt_xdate()
fig.tight_layout()
fig.savefig("revenue_trend.png", dpi=200)
plt.show()

Read the diff

One faded line shows the actual daily revenue (noisy); a bold line shows the trend. The Monday-locator and dd-Mon formatter keep the date axis readable. PNG saved at 200 DPI — looks crisp on a slide.

Try It Yourself

13 min

01 🟢 Simplest line

Plot any column from a time-series CSV against its date column. Add a title and y-label.

02 🟡 Two series

Plot two metrics on the same axes with a legend (e.g., open vs done counts, daily vs weekly revenue).

03 🔴 Twin axes

When two series have very different scales, use ax2 = ax.twinx() to give them their own y-axes.

Hint

fig, ax = plt.subplots()
ax.plot(dates, revenue, color="b", label="revenue")
ax.set_ylabel("Revenue (RM)", color="b")

ax2 = ax.twinx()
ax2.plot(dates, orders, color="r", label="orders")
ax2.set_ylabel("Orders", color="r")

Mini-Challenge · Compare Years

8 min

From a daily dataset spanning ≥ 2 years, plot one line per year, x-axis = day-of-year (1-365). The result lets you eyeball seasonal differences.

Show one possible solution

df["doy"]  = df["date"].dt.dayofyear
df["year"] = df["date"].dt.year

fig, ax = plt.subplots(figsize=(10, 5))
for yr, g in df.groupby("year"):
    ax.plot(g["doy"], g["sales"].rolling(7).mean(), label=str(yr))
ax.legend()
ax.set_title("Daily sales — 7-day avg by year")
ax.set_xlabel("Day of year")

Rolling averages let multi-year overlays be readable. Without smoothing, four lines on top of each other become noise.

Recap

3 min

ax.plot for lines. Multiple series + legend for comparison. Date axes via matplotlib.dates. A 7-day rolling average tames noisy daily data. Save as PNG. Tomorrow: pie + scatter for proportions and correlations.

Homework

4 min

Draw a time-series chart from your real CSV that includes a 7-day rolling average overlay, a date axis formatted readably, and an annotation marking the peak point. Save the PNG.

import matplotlib.dates as mdates import pandas as pd df = pd.read_csv("daily.csv", parse_dates=["date"]) fig, ax = plt.subplots(figsize=(10, 5)) ax.plot(df["date"], df["sales"]) # major tick every month, format as "May 2026" ax.xaxis.set_major_locator(mdates.MonthLocator()) ax.xaxis.set_major_formatter(mdates.DateFormatter("%b %Y")) fig.autofmt_xdate()

peak = df.loc[df["sales"].idxmax()] ax.annotate(f"peak: {int(peak['sales'])}", xy=(peak["date"], peak["sales"]), xytext=(peak["date"], peak["sales"] + 30), arrowprops={"arrowstyle": "->"})

# daily_revenue.py import pandas as pd import matplotlib.pyplot as plt import matplotlib.dates as mdates df = pd.read_csv("clean.csv", parse_dates=["date"]) df["total"] = df["quantity"] * df["price"] daily = (df.groupby(df["date"].dt.date)["total"].sum() .rename("daily").to_frame()) daily.index = pd.to_datetime(daily.index) daily["roll7"] = daily["daily"].rolling(7).mean() fig, ax = plt.subplots(figsize=(10, 5)) ax.plot(daily.index, daily["daily"], color="#4a90e2", alpha=0.35, label="daily") ax.plot(daily.index, daily["roll7"], color="#1f5fa8", linewidth=2.5, label="7-day avg") ax.set_title("Daily revenue · 7-day rolling") ax.set_ylabel("RM") ax.xaxis.set_major_locator(mdates.WeekdayLocator(byweekday=mdates.MO)) ax.xaxis.set_major_formatter(mdates.DateFormatter("%d %b")) ax.grid(axis="y", linestyle="--", alpha=0.4) ax.legend() fig.autofmt_xdate() fig.tight_layout() fig.savefig("revenue_trend.png", dpi=200) plt.show()

fig, ax = plt.subplots() ax.plot(dates, revenue, color="b", label="revenue") ax.set_ylabel("Revenue (RM)", color="b") ax2 = ax.twinx() ax2.plot(dates, orders, color="r", label="orders") ax2.set_ylabel("Orders", color="r")

df["doy"] = df["date"].dt.dayofyear df["year"] = df["date"].dt.year fig, ax = plt.subplots(figsize=(10, 5)) for yr, g in df.groupby("year"): ax.plot(g["doy"], g["sales"].rolling(7).mean(), label=str(yr)) ax.legend() ax.set_title("Daily sales — 7-day avg by year") ax.set_xlabel("Day of year")