import matplotlib.pyplot as plt
fig, ax = plt.subplots(figsize=(8, 4))
df[“value”].hist(ax=ax, bins=30)
ax.set_title(“Value distribution”)
ax.set_xlabel(“value”)
ax.set_ylabel(“count”)
fig.tight_layout()
fig.savefig(“figures/value_distribution.png”, dpi=160)

week04-visualization-eda/
├── data/
│   ├── raw/
│   │   └── student_scores.csv
│   └── processed/
│       └── student_scores_clean.csv
├── figures/
│   ├── average_score_hist.png
│   ├── study_hours_vs_average_score.png
│   ├── average_score_by_major.png
│   └── correlation_heatmap.png
├── notebooks/
│   └── week04_visualization_eda.ipynb
├── reports/
│   └── week04_eda_report.md
├── src/
│   ├── prepare_data.py
│   └── make_figures.py
├── .venv/
├── pyproject.toml
└── README.md

mkdir -p ~/Code/python-learning/week04-visualization-eda
cd ~/Code/python-learning/week04-visualization-eda
mkdir -p src data/raw data/processed figures notebooks reports
uv init
uv venv
source .venv/bin/activate.fish
uv add numpy pandas matplotlib
code .

uv run python -c "import pandas, matplotlib; print('ok')"

uv add seaborn

printf 'student,major,gender,math,english,python,study_hours\nA,statistics,F,88,79,85,12\nB,statistics,M,92,81,90,14\nC,finance,F,75,86,78,9\nD,finance,M,80,,82,8\nE,statistics,F,95,88,,15\nF,computer,M,70,76,88,10\nG,computer,F,,90,91,11\nH,statistics,M,84,82,86,13\nI,finance,F,78,84,80,9\nJ,computer,M,73,78,84,10\nK,statistics,F,90,85,92,14\nL,finance,M,82,80,79,7\n' > data/raw/student_scores.csv

head data/raw/student_scores.csv
wc -l data/raw/student_scores.csv

from pathlib import Path
import pandas as pd
RAW_PATH = Path(“data/raw/student_scores.csv”)
OUTPUT_PATH = Path(“data/processed/student_scores_clean.csv”)
NUMERIC_COLUMNS = [“math”, “english”, “python”, “study_hours”]
SCORE_COLUMNS = [“math”, “english”, “python”]
def load_data(path: Path) -> pd.DataFrame:
return pd.read_csv(path)
def clean_data(df: pd.DataFrame) -> pd.DataFrame:
cleaned = df.copy()
plain for column in NUMERIC_COLUMNS:
cleaned[column] = pd.to_numeric(cleaned[column], errors=“coerce”)
cleaned[column] = cleaned[column].fillna(cleaned[column].median())
plain cleaned[“average_score”] = cleaned[SCORE_COLUMNS].mean(axis=1)
return cleaned
def save_data(df: pd.DataFrame, path: Path) -> None:
path.parent.mkdir(parents=True, exist_ok=True)
df.to_csv(path, index=False)
def main() -> None:
raw = load_data(RAW_PATH)
cleaned = clean_data(raw)
save_data(cleaned, OUTPUT_PATH)
print(f”Rows: {len(cleaned)}”)
print(f”Columns: {len(cleaned.columns)}”)
print(f”Saved: {OUTPUT_PATH}”)
if name == “main”:
main()

uv run python src/prepare_data.py
ls data/processed
head data/processed/student_scores_clean.csv

import matplotlib.pyplot as plt
fig, ax = plt.subplots(figsize=(8, 5))
ax.hist(values, bins=10)
ax.set_title(“Title”)
ax.set_xlabel(“X label”)
ax.set_ylabel(“Y label”)
fig.tight_layout()
fig.savefig(“figures/example.png”, dpi=150)
plt.close(fig)

from pathlib import Path
import matplotlib.pyplot as plt
import pandas as pd
DATA_PATH = Path(“data/processed/student_scores_clean.csv”)
FIGURE_DIR = Path(“figures”)
SCORE_COLUMNS = [“math”, “english”, “python”, “average_score”]
def load_data(path: Path) -> pd.DataFrame:
return pd.read_csv(path)
def save_current_figure(fig, filename: str) -> None:
FIGURE_DIR.mkdir(parents=True, exist_ok=True)
output_path = FIGURE_DIR / filename
fig.tight_layout()
fig.savefig(output_path, dpi=150)
plt.close(fig)
print(f”Saved {output_path}”)
def plot_average_score_hist(df: pd.DataFrame) -> None:
fig, ax = plt.subplots(figsize=(8, 5))
ax.hist(df[“average_score”], bins=8, edgecolor=“black”)
ax.set_title(“Distribution of Average Score”)
ax.set_xlabel(“Average score”)
ax.set_ylabel(“Student count”)
save_current_figure(fig, “average_score_hist.png”)
def plot_study_hours_scatter(df: pd.DataFrame) -> None:
fig, ax = plt.subplots(figsize=(8, 5))
ax.scatter(df[“study_hours”], df[“average_score”])
ax.set_title(“Study Hours vs Average Score”)
ax.set_xlabel(“Study hours per week”)
ax.set_ylabel(“Average score”)
save_current_figure(fig, “study_hours_vs_average_score.png”)
def plot_major_boxplot(df: pd.DataFrame) -> None:
majors = sorted(df[“major”].unique())
values = [df.loc[df[“major”] == major, “average_score”] for major in majors]
plain fig, ax = plt.subplots(figsize=(8, 5))
ax.boxplot(values, labels=majors)
ax.set_title(“Average Score by Major”)
ax.set_xlabel(“Major”)
ax.set_ylabel(“Average score”)
save_current_figure(fig, “average_score_by_major.png”)
def plot_correlation_heatmap(df: pd.DataFrame) -> None:
corr = df[SCORE_COLUMNS + [“study_hours”]].corr()
plain fig, ax = plt.subplots(figsize=(7, 6))
image = ax.imshow(corr, cmap=“coolwarm”, vmin=-1, vmax=1)
plain ax.set_xticks(range(len(corr.columns)))
ax.set_yticks(range(len(corr.index)))
ax.set_xticklabels(corr.columns, rotation=45, ha=“right”)
ax.set_yticklabels(corr.index)
ax.set_title(“Correlation Heatmap”)
plain for row in range(len(corr.index)):
for col in range(len(corr.columns)):
ax.text(col, row, f”{corr.iloc[row, col]:.2f}”, ha=“center”, va=“center”, color=“black”)
fig.colorbar(image, ax=ax)
save_current_figure(fig, "correlation_heatmap.png")
def main() -> None:
df = load_data(DATA_PATH)
plot_average_score_hist(df)
plot_study_hours_scatter(df)
plot_major_boxplot(df)
plot_correlation_heatmap(df)
if name == “main”:
main()

uv run python src/prepare_data.py
uv run python src/make_figures.py
ls figures

average_score_by_major.png
average_score_hist.png
correlation_heatmap.png
study_hours_vs_average_score.png

ax.hist(df["average_score"], bins=8, edgecolor="black")

平均分主要集中在 80 到 90 分之间，样本中没有特别低的平均分。由于样本量很小，这个分布只能用于流程练习，不能代表真实学生群体。

ax.scatter(df["study_hours"], df["average_score"])

散点图显示学习时长和平均分大致同向变化：学习时长较高的学生平均分往往也较高。但样本只有 12 行，不能据此得出稳定因果结论。

majors = sorted(df["major"].unique())
values = [df.loc[df["major"] == major, "average_score"] for major in majors]
ax.boxplot(values, labels=majors)

statistics 组平均分中位数较高，computer 和 finance 组略低。但每组样本数量很少，箱线图主要用于展示方法，不适合做群体差异结论。

corr = df[["math", "english", "python", "average_score", "study_hours"]].corr()

uv add seaborn

import seaborn as sns
corr = df[[“math”, “english”, “python”, “average_score”, “study_hours”]].corr()
fig, ax = plt.subplots(figsize=(7, 6))
sns.heatmap(corr, annot=True, cmap=“coolwarm”, vmin=-1, vmax=1, ax=ax)
fig.tight_layout()
fig.savefig(“figures/correlation_heatmap_seaborn.png”, dpi=150)
plt.close(fig)

ls figures

1.png
new.png
final_final.png

# Week 04 EDA 报告
1. 数据来源
说明数据来自本周练习 CSV，包含学生专业、性别、三科成绩和每周学习时长。
2. 字段说明


















































字段
含义
类型
student
学生编号
分类
major
专业
分类
gender
性别
分类
math
数学成绩
数值
english
英语成绩
数值
python
Python 成绩
数值
study_hours
每周学习时长
数值
average_score
三科平均分
数值
3. 缺失值情况
说明原始数据中哪些列有缺失值，以及使用中位数填补数值列。
4. 主要分布
引用 figures/average_score_hist.png，描述平均分集中区间和异常值情况。
5. 分组观察
引用 figures/average_score_by_major.png，描述不同专业组的平均分分布，并说明样本量限制。
6. 相关性观察
引用 figures/study_hours_vs_average_score.png 和 figures/correlation_heatmap.png，描述学习时长与平均分、各科成绩之间的相关性。
7. 可能的建模目标
可以提出一个后续建模问题，例如：根据学习时长、专业和前两科成绩预测 Python 成绩，或根据特征判断 average_score 是否高于 85。
8. 局限性
说明样本量小、数据为练习数据、不能代表真实 USTC 学生群体。

![平均分直方图](../figures/average_score_hist.png)

notebooks/week04_visualization_eda.ipynb：探索
src/prepare_data.py：固定清洗流程
src/make_figures.py：固定画图流程
reports/week04_eda_report.md：写最终解释

uv add ipykernel

source .venv/bin/activate.fish
uv run python src/prepare_data.py
uv run python src/make_figures.py
ls figures

uv sync
uv run python src/prepare_data.py
uv run python src/make_figures.py

def plot_python_score_hist(df: pd.DataFrame) -> None:
    fig, ax = plt.subplots(figsize=(8, 5))
    ax.hist(df["python"], bins=8, edgecolor="black")
    ax.set_title("Distribution of Python Score")
    ax.set_xlabel("Python score")
    ax.set_ylabel("Student count")
    save_current_figure(fig, "python_score_hist.png")

uv run python src/make_figures.py
ls figures

average_score_by_gender.png

from pathlib import Path
import pandas as pd
df = pd.read_csv(Path(“data/processed/student_scores_clean.csv”))
summary = df.describe()
summary.to_csv(Path(“reports/descriptive_statistics.csv”))

uv run python src/export_summary.py
ls reports

pwd
ls
source .venv/bin/activate.fish
uv run python src/prepare_data.py
uv run python src/make_figures.py
ls data/processed
ls figures

cd ~/Code/python-learning/week04-visualization-eda
uv add matplotlib
uv run python src/make_figures.py

uv run python src/make_figures.py
ls figures

fig.savefig(output_path, dpi=150)

FIGURE_DIR.mkdir(parents=True, exist_ok=True)

uv run python src/prepare_data.py
uv run python src/make_figures.py

ax.imshow(corr, cmap="coolwarm", vmin=-1, vmax=1)