Data: the fuel of AI.

An AI model is only as good as the data it learns from. Today you'll learn what data looks like, how to explore it with Python's pandas library, and why "garbage in, garbage out" is the first rule of AI.

Data as a table

Most AI data starts as a table: rows are examples, columns are details.

• Features: the inputs (area, bedrooms, locality).
• Label / target: what we want to predict (price).

Garbage in, garbage out

If the data is wrong, biased or messy, the model learns the wrong things.

Exploring data with pandas

In Google Colab:

import pandas as pd

data = {
    "area_sqft": [1200, 1650, 900, 1400, 2000],
    "bedrooms": [2, 3, 2, 3, 3],
    "price_lakh": [72, 118, 55, 90, 140],
}
df = pd.DataFrame(data)

print(df.head())          # first rows
print(df.describe())      # averages, min, max
print(df["price_lakh"].mean())   # average price
print(df[df["bedrooms"] == 3])   # only 3-bedroom flats

A DataFrame is pandas' version of a spreadsheet. You can sort, filter and calculate just like in Excel, but faster and with millions of rows.

Loading real data

df = pd.read_csv("your_file.csv")
df.info()           # column types and missing values
df.isnull().sum()   # count missing values per column
df = df.drop_duplicates()

You can upload a CSV in Colab using the folder icon on the left. Free datasets: Kaggle, data.gov.in (Indian government open data).

Privacy matters

Real data often contains people's personal information. Never share or upload private customer data to unknown tools, and follow your company's rules. Practise with public datasets.