Artificial Intelligence Basics for Children in Taichung City

1. Beyond Averages

Means alone hide spread. Error bars on bars and confidence bands on lines show how much values vary around a central estimate.

2. Error Bars on Category Means

import pandas as pd
import matplotlib.pyplot as plt

df = pd.read_csv("ecommerce_orders.csv")
stats = df.groupby("Product Category")["Revenue"].agg(["mean", "std"])

fig, ax = plt.subplots(figsize=(8, 5))
ax.bar(
    stats.index,
    stats["mean"],
    yerr=stats["std"],
    capsize=5,
    color="skyblue",
    edgecolor="black"
)
ax.set_title("Mean Revenue by Product Category with Error Bars (Std Dev)")
ax.set_xlabel("Product Category")
ax.set_ylabel("Mean Revenue ($)")
plt.tight_layout()
plt.show()

yerr draws vertical whiskers; capsize sets the width of the end caps.

3. Confidence Band on Monthly Revenue

df_sample = df.head(500).copy()
df_sample["Order Date"] = pd.to_datetime(df_sample["Order Date"])
df_sample = df_sample.set_index("Order Date")
monthly = df_sample["Revenue"].resample("ME").agg(["mean", "std"])

fig, ax = plt.subplots(figsize=(10, 6))
ax.plot(monthly.index, monthly["mean"], color="blue", label="Mean Revenue")
ax.fill_between(
    monthly.index,
    monthly["mean"] - monthly["std"],
    monthly["mean"] + monthly["std"],
    color="blue",
    alpha=0.3,
    label="±1 Std Dev Confidence Band"
)
ax.tick_params(axis="x", labelrotation=45)
ax.set_title("Monthly Mean Revenue with Confidence Band (±1 Std Dev)")
ax.legend()
plt.tight_layout()
plt.show()

fill_between shades mean±std; alpha=0.3 keeps the line readable.

4. Parameter Cheat Sheet

Parameter	Role
yerr	Error whiskers on bars
capsize	Horizontal cap width
fill_between	Shaded band between bounds
alpha	Band transparency
labelrotation	Angle x labels (e.g. 45°)