Exploratory data analysis and charts

import numpy as np
import pandas as pd

rng = np.random.default_rng(42)
n = 365
dates = pd.date_range("2025-11-01", periods=n, freq="D")
temp = 24 + 6 * np.sin((dates.day_of_year.to_numpy() - 80) / 365 * 2 * np.pi) + rng.normal(0, 2, n)
weekend = (dates.day_of_week >= 5).astype(int)
promo = rng.binomial(1, 0.2, n)
cups = 90 + 6 * (temp - 24) + 25 * weekend + 18 * promo + rng.normal(0, 10, n)

sales = pd.DataFrame({
    "date": dates,
    "temp_c": temp.round(1),
    "is_weekend": weekend,
    "promo": promo,
    "cups_sold": cups.round().astype(int),
})
# Make it messy, like real data: 8 missing temperatures and one wrong entry
sales.loc[rng.choice(n, 8, replace=False), "temp_c"] = np.nan
sales.loc[100, "cups_sold"] = 900
sales.to_csv("sales.csv", index=False)
print("saved sales.csv with", len(sales), "rows")

import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns

sales = pd.read_csv("sales.csv", parse_dates=["date"])
print(sales.shape)
print(sales.head())
print(sales.isna().sum())
print(sales.describe().round(1))

q1, q3 = sales["cups_sold"].quantile([0.25, 0.75])
iqr = q3 - q1
low, high = q1 - 1.5 * iqr, q3 + 1.5 * iqr
print("normal range:", low, "to", high)
print(sales[(sales["cups_sold"] < low) | (sales["cups_sold"] > high)])

clean = sales[sales["cups_sold"] < 500].copy()
clean["temp_c"] = clean["temp_c"].fillna(clean["temp_c"].median())
print(len(clean), clean.isna().sum().sum())

fig, axes = plt.subplots(1, 2, figsize=(10, 4))

axes[0].hist(clean["cups_sold"], bins=20, color="#4C78A8", edgecolor="white")
axes[0].set_title("Cups sold per day")
axes[0].set_xlabel("Cups")
axes[0].set_ylabel("Number of days")

sns.boxplot(data=clean, x="is_weekend", y="cups_sold", ax=axes[1])
axes[1].set_title("Cups sold: weekday (0) vs weekend (1)")

plt.tight_layout()
plt.show()

sns.scatterplot(data=clean, x="temp_c", y="cups_sold", hue="promo", alpha=0.7)
plt.title("Hotter days sell more; promo days sell more at the same temperature")
plt.xlabel("Temperature (deg C)")
plt.ylabel("Cups sold")
plt.show()

print(clean.groupby("is_weekend")["cups_sold"].mean().round(1))
print(clean.groupby("promo")["cups_sold"].mean().round(1))

cols = ["temp_c", "is_weekend", "promo", "cups_sold"]
corr = clean[cols].corr().round(2)
print(corr)

sns.heatmap(corr, annot=True, cmap="coolwarm", vmin=-1, vmax=1)
plt.title("Correlation between columns")
plt.show()

monthly = clean.set_index("date")["cups_sold"].resample("MS").mean()
monthly.plot(kind="line", marker="o", figsize=(9, 3.5))
plt.title("Average cups per day, by month")
plt.ylabel("Cups")
plt.xlabel("Month")
plt.show()

clean["weekday"] = clean["date"].dt.day_name()
order = ["Monday", "Tuesday", "Wednesday", "Thursday", "Friday", "Saturday", "Sunday"]
clean.groupby("weekday")["cups_sold"].mean().reindex(order).plot(kind="bar", figsize=(8, 3.5))
plt.title("Average cups by day of week")
plt.ylabel("Cups")
plt.show()

# Findings (written in a notebook text cell)
# 1. 365 rows; 8 missing temperatures filled with median; one 900-cup entry removed (billing error).
# 2. Sales swing with temperature: correlation about 0.85.
# 3. Weekends average about 25 cups more than weekdays; promo days about 20 more.
# 4. Next step: build a regression model using temp_c, is_weekend and promo.

I explored a dataset in pandas. Here is what I found:

[paste describe() output and your findings]

Act as a senior data scientist. Tell me: (1) what I may have missed in data quality, (2) which charts I should add, (3) any claims I made that could be wrong because of correlation vs causation, and (4) what I should check before building a model. Keep it to a short list.