Don't miss anomalies with a control chart

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37

import numpy as np
import matplotlib.pyplot as plt

rng = np.random.default_rng(5)
values = 45 + rng.normal(0, 3, size=28)
values[[6, 18]] += np.array([12, -10])  # Inject anomalies

mean = values.mean()
std = values.std(ddof=1)
ucl = mean + 3 * std
lcl = mean - 3 * std

fig, ax = plt.subplots(figsize=(6.4, 3.6))
ax.plot(values, marker="o", color="#0ea5e9")
ax.axhline(mean, color="#334155", linewidth=1.3, label="Mean")
ax.axhline(ucl, color="#ef4444", linestyle="--", label="UCL")
ax.axhline(lcl, color="#ef4444", linestyle="--", label="LCL")

ax.set_xticks(range(0, len(values), 4), labels=[f"W{i+1}" for i in range(0, len(values), 4)])
ax.set_title("Control chart of call handling time")
ax.set_ylabel("Average handling time (sec)")
ax.grid(alpha=0.2)

for idx, val in enumerate(values):
    if val > ucl or val < lcl:
        ax.annotate(
            "Anomaly",
            (idx, val),
            xytext=(idx + 0.5, val + 4),
            arrowprops=dict(arrowstyle="->", color="#ef4444"),
            color="#ef4444",
        )

ax.legend(loc="upper right")
fig.tight_layout()

plt.show()