On Day 26 you expand beyond data wrangling and apply core statistical tools to business datasets. The refactored lesson script now exposes reusable helper functions for descriptive statistics, correlation analysis, and hypothesis testing so you can integrate them directly into your own notebooks or projects.
Before you begin, ensure you have followed the setup instructions in the main README.md to create your virtual environment and install the required libraries.
The stats.py module has been organized into testable functions that separate
computation from presentation.
Day_26_Statistics/stats.py and look at the new
helpers:
load_sales_data() reads sales_data.csv and removes missing rows.summarize_revenue() returns key revenue metrics and the full
DataFrame.describe() output.compute_correlations() produces a correlation matrix for the numeric
sales fields.run_ab_test() wraps SciPy’s independent t-test and reports whether the
difference is statistically significant.build_revenue_distribution_chart() visualises the revenue histogram as an
interactive Plotly figure.build_correlation_heatmap() turns the correlation matrix into an
interactive heatmap with hover labels and colourbar explanations.python Day_26_Statistics/stats.py
pytest tests/test_day_26.py
Static Matplotlib previews are great for reports, but sometimes you need to hover over exact values or export a filtered view. Open the companion notebook to try the Plotly charts side by side with their static counterparts:
pip install notebook plotly
jupyter notebook Day_26_Statistics/statistics_interactive.ipynb
🎉 Great job! With these reusable statistics utilities you can move from simple summaries to rigorous, testable insights in your analyses.
Run this lesson’s code interactively in your browser:
!!! tip “About JupyterLite” JupyterLite runs entirely in your browser using WebAssembly. No installation or server required! Note: First launch may take a moment to load.
???+ example “solutions.py” View on GitHub
```python title="solutions.py"
"""
Day 26: Solutions to Exercises
"""
from pathlib import Path
import pandas as pd
from scipy.stats import ttest_ind
# --- Exercise 1: Descriptive Statistics of Sales ---
print("--- Solution to Exercise 1 ---")
resource_dir = Path(__file__).resolve().parent
data_path = resource_dir / "sales_data.csv"
try:
df = pd.read_csv(data_path)
df.dropna(inplace=True) # Clean the data first
revenue = df["Revenue"]
print(f"Mean Revenue: ${revenue.mean():,.2f}")
print(f"Median Revenue: ${revenue.median():,.2f}")
print(f"Standard Deviation of Revenue: ${revenue.std():,.2f}")
print(f"Minimum Revenue: ${revenue.min():,.2f}")
print(f"Maximum Revenue: ${revenue.max():,.2f}")
except FileNotFoundError:
print("Error: sales_data.csv not found. Keep the CSV beside this script.")
df = pd.DataFrame()
print("-" * 20)
# --- Exercise 2: Correlation Analysis ---
print("--- Solution to Exercise 2 ---")
if not df.empty:
# Select only the numerical columns
numerical_cols = df[["Units Sold", "Price", "Revenue"]]
# Calculate the correlation matrix
correlation_matrix = numerical_cols.corr()
print("Correlation Matrix:")
print(correlation_matrix)
print(
"\nAnswer: 'Units Sold' and 'Revenue' have the strongest positive correlation (0.93)."
)
else:
print("DataFrame not available for this exercise.")
print("-" * 20)
# --- Exercise 3: A/B Test Analysis (T-Test) ---
print("--- Solution to Exercise 3 ---")
group_a_durations = [10.5, 12.1, 11.8, 13.0, 12.5]
group_b_durations = [12.8, 13.5, 13.2, 14.0, 13.8]
print(f"Group A Durations: {group_a_durations}")
print(f"Group B Durations: {group_b_durations}")
# Perform the t-test
t_stat, p_value = ttest_ind(group_a_durations, group_b_durations)
print(f"\nP-value: {p_value:.4f}")
# Conclusion based on the p-value
if p_value < 0.05: # pyright: ignore[reportOperatorIssue]
print(
"Conclusion: The result is statistically significant. The two headlines likely have different effects on session duration."
)
else:
print(
"Conclusion: The result is not statistically significant. We cannot conclude there is a difference between the headlines."
)
print("-" * 20)
```
???+ example “stats.py” View on GitHub
```python title="stats.py"
"""Day 26: Practical Statistics in Python.
This module provides helper functions for loading sales data, generating
summary statistics, computing correlations, and running a simple A/B test.
All side effects (such as printing to the console) are encapsulated in the
``main`` function so that the individual helpers are easy to import and test.
"""
from __future__ import annotations
from pathlib import Path
from typing import Iterable, Mapping, MutableMapping
import pandas as pd
import plotly.graph_objects as go
from pandas import DataFrame, Series
from scipy.stats import ttest_ind
def load_sales_data(csv_path: Path | str | None = None) -> DataFrame:
"""Load and clean the sales CSV file."""
resource_dir = Path(__file__).resolve().parent
path = Path(csv_path) if csv_path is not None else resource_dir / "sales_data.csv"
try:
df = pd.read_csv(path)
except FileNotFoundError:
return pd.DataFrame()
return df.dropna(axis=0, how="any")
def summarize_revenue(df: DataFrame) -> Mapping[str, float | Series | DataFrame]:
"""Return descriptive statistics for the ``Revenue`` column."""
if "Revenue" not in df:
raise KeyError("DataFrame must contain a 'Revenue' column")
revenue = df["Revenue"]
summary: MutableMapping[str, float | Series | DataFrame] = {
"mean": float(revenue.mean()),
"median": float(revenue.median()),
"std": float(revenue.std()),
"min": float(revenue.min()),
"max": float(revenue.max()),
"describe": df.describe(),
}
return summary
def compute_correlations(df: DataFrame) -> DataFrame:
"""Return the correlation matrix for the key numeric columns."""
columns = [col for col in ("Units Sold", "Price", "Revenue") if col in df.columns]
if len(columns) < 2:
raise ValueError(
"At least two of 'Units Sold', 'Price', or 'Revenue' must be present"
)
return df[columns].corr()
def build_revenue_distribution_chart(df: DataFrame) -> go.Figure:
"""Create a histogram visualising the distribution of the ``Revenue`` column."""
if "Revenue" not in df:
raise KeyError("DataFrame must contain a 'Revenue' column")
revenue = df["Revenue"].dropna()
figure = go.Figure(
data=[
go.Histogram(
x=revenue,
nbinsx=min(30, max(5, revenue.nunique() // 2 or 5)),
marker_color="#636EFA",
opacity=0.85,
hovertemplate="Revenue: %{x:$,.0f}<extra></extra>",
)
]
)
figure.update_layout(
title="Revenue Distribution",
xaxis_title="Revenue",
yaxis_title="Frequency",
template="plotly_white",
bargap=0.05,
)
return figure
def build_correlation_heatmap(df: DataFrame) -> go.Figure:
"""Create a heatmap to visualise correlations between key numeric metrics."""
correlations = compute_correlations(df)
heatmap = go.Heatmap(
z=correlations.values,
x=list(correlations.columns),
y=list(correlations.index),
colorscale="Blues",
zmin=-1,
zmax=1,
hovertemplate="%{y} vs %{x}: %{z:.2f}<extra></extra>",
colorbar=dict(title="Correlation"),
)
figure = go.Figure(data=[heatmap])
figure.update_layout(
title="Correlation Heatmap",
template="plotly_white",
)
return figure
def run_ab_test(
group_a: Iterable[float], group_b: Iterable[float], alpha: float = 0.05
) -> Mapping[str, float | bool]:
"""Run an independent t-test on two groups of durations."""
t_statistic, p_value = ttest_ind(list(group_a), list(group_b))
return {
"t_statistic": float(t_statistic),
"p_value": float(p_value),
"alpha": float(alpha),
"is_significant": bool(p_value < alpha),
}
def main() -> None:
"""Execute the lesson workflow with helpful console output."""
print("--- 1. Descriptive Statistics of Sales Data ---")
df = load_sales_data()
if df.empty:
print("Error: sales_data.csv not found. Keep the CSV beside this script.")
else:
revenue_summary = summarize_revenue(df)
print(f"Mean Revenue: ${revenue_summary['mean']:,.2f}")
print(f"Median Revenue: ${revenue_summary['median']:,.2f}")
print(f"Standard Deviation of Revenue: ${revenue_summary['std']:,.2f}")
print(f"Minimum Revenue: ${revenue_summary['min']:,.2f}")
print(f"Maximum Revenue: ${revenue_summary['max']:,.2f}")
print("\nFull descriptive statistics (df.describe()):")
print(revenue_summary["describe"])
print("-" * 20)
print("--- 2. Correlation Analysis ---")
if not df.empty:
correlation_matrix = compute_correlations(df)
print("Correlation Matrix:")
print(correlation_matrix)
print(
"\nAnalysis: 'Units Sold' and 'Revenue' have a strong positive correlation (0.93)."
)
print("'Price' and 'Revenue' also have a strong positive correlation (0.83).")
print(
"'Price' and 'Units Sold' have a weak negative correlation (-0.23), which might be expected (higher price can sometimes mean fewer units)."
)
else:
print("DataFrame not available for this exercise.")
print("-" * 20)
print("--- 3. A/B Test Analysis (T-Test) ---")
group_a_durations = [10.5, 12.1, 11.8, 13.0, 12.5, 11.9, 12.3]
group_b_durations = [12.8, 13.5, 13.2, 14.0, 13.8, 14.1, 13.6]
print(f"Group A (Old Headline) Durations: {group_a_durations}")
print(f"Group B (New Headline) Durations: {group_b_durations}")
test_results = run_ab_test(group_a_durations, group_b_durations)
print(f"\nT-statistic: {test_results['t_statistic']:.4f}")
print(f"P-value: {test_results['p_value']:.4f}")
if test_results["is_significant"]:
print("\nConclusion: The difference is statistically significant.")
print(
"We can conclude that the new headline (Group B) likely leads to longer session durations."
)
else:
print("\nConclusion: The difference is not statistically significant.")
print(
"We cannot conclude that the new headline had a real effect on session duration."
)
print("-" * 20)
if __name__ == "__main__":
main()
```