दिन 18: Exploratory Data Analysis (EDA)

import pandas as pd
h = pd.read_csv("house_prices.csv")
print(h.shape)
print(h.head(3))
print("Missing:", h.isna().sum().sum(), "| duplicates:", h.duplicated().sum())
print(h.describe().round(1).T[["mean", "50%", "min", "max"]])

import pandas as pd
import matplotlib.pyplot as plt
h = pd.read_csv("house_prices.csv")
fig, ax = plt.subplots(1, 3, figsize=(12, 3.6))
ax[0].hist(h["price_lakh"], bins=30, color="#4f46e5"); ax[0].set_title("Price (₹ lakh)")
ax[1].hist(h["area_sqft"], bins=30, color="#0891b2"); ax[1].set_title("Area (sqft)")
h["city"].value_counts().plot(kind="bar", ax=ax[2], color="#f59e0b"); ax[2].set_title("Listings per city")
ax[2].tick_params(axis="x", rotation=30)
plt.tight_layout(); plt.show()
print("Skew → price:", round(h["price_lakh"].skew(), 2), "| area:", round(h["area_sqft"].skew(), 2))

import pandas as pd
h = pd.read_csv("house_prices.csv")
q1, q3 = h["price_lakh"].quantile([0.25, 0.75])
iqr = q3 - q1
low, high = q1 - 1.5 * iqr, q3 + 1.5 * iqr
out = h[(h["price_lakh"] < low) | (h["price_lakh"] > high)]
print(f"Q1={q1:.1f} Q3={q3:.1f} IQR={iqr:.1f} → normal range {low:.1f} to {high:.1f} lakh")
print(len(out), "outliers")
print(out.sort_values("price_lakh", ascending=False)[["city", "area_sqft", "bedrooms", "price_lakh"]].head())

import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
h = pd.read_csv("house_prices.csv")
order = h.groupby("city")["price_lakh"].median().sort_values().index
plt.figure(figsize=(9, 4))
sns.boxplot(data=h, x="city", y="price_lakh", order=order, color="#a5b4fc")
plt.title("Price by city – boxes show the middle 50%, dots are outliers")
plt.ylabel("₹ lakh"); plt.show()

import pandas as pd
import matplotlib.pyplot as plt
h = pd.read_csv("house_prices.csv")
fig, ax = plt.subplots(1, 2, figsize=(11, 4))
ax[0].scatter(h["area_sqft"], h["price_lakh"], s=10, alpha=0.5, color="#4f46e5")
ax[0].set_xlabel("Area (sqft)"); ax[0].set_ylabel("Price (₹ lakh)"); ax[0].set_title("Bigger homes cost more")
h["ppsf"] = h["price_lakh"] * 1e5 / h["area_sqft"]               # price per sqft removes the size effect
band = pd.cut(h["metro_km"], [0, 1, 2, 4, 7, 18], labels=["<1", "1-2", "2-4", "4-7", "7+"])
h.groupby(band, observed=True)["ppsf"].median().plot(kind="bar", ax=ax[1], color="#dc2626", rot=0)
ax[1].set_xlabel("Distance to metro (km)"); ax[1].set_ylabel("₹ per sqft"); ax[1].set_title("Farther from metro → cheaper per sqft")
plt.tight_layout(); plt.show()

import pandas as pd
h = pd.read_csv("house_prices.csv")
for col in ["furnishing", "parking", "bedrooms"]:
    print(h.groupby(col)["price_lakh"].median().round(1).to_dict())

import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
h = pd.read_csv("house_prices.csv")
corr = h.select_dtypes("number").corr().round(2)
print(corr["price_lakh"].sort_values(ascending=False))
plt.figure(figsize=(6.5, 5))
sns.heatmap(corr, annot=True, cmap="RdBu_r", vmin=-1, vmax=1, fmt=".2f")
plt.title("Correlation matrix"); plt.show()

import pandas as pd
def quick_eda(df, target):
    print("Shape:", df.shape, "| missing:", int(df.isna().sum().sum()), "| dups:", int(df.duplicated().sum()))
    num = df.select_dtypes("number").drop(columns=[target])
    print("Top numeric drivers:", num.corrwith(df[target]).abs().sort_values(ascending=False).round(2).head(3).to_dict())
    for c in df.select_dtypes("object").columns:
        g = df.groupby(c)[target].median()
        print(f"{c}: best={g.idxmax()} ({g.max():.1f}) worst={g.idxmin()} ({g.min():.1f})")
quick_eda(pd.read_csv("house_prices.csv"), "price_lakh")