Pandas Basics for Interviews: Load, Select and Filter Data with Confidence

A category manager opens a CSV with 12 lakh order rows five minutes before a Monday review. The business question is simple - “Which premium skincare SKUs underperformed in South India last week?” - but the answer appears only after the noise is loaded, sliced and filtered correctly.

Big Picture: Pandas Is the Analyst’s Sieve

Pandas is not “just Python syntax.” It is the tool that lets a manager move from a raw business dump to the exact slice needed for a decision. Loading brings the file into memory; selecting chooses the variables; filtering chooses the observations.

Core Explanation: The Three Skills You Must Control

The core idea is simple: a Pandas DataFrame behaves like an Excel sheet with far more power. Rows usually represent observations such as orders, customers or transactions. Columns represent variables such as date, city, category, price or revenue.

1. Loading Data: Bring the Source into a DataFrame

The most common convention is:

import pandas as pd
df = pd.read_csv("orders.csv")

Here, pd is the alias for Pandas, read_csv() reads a comma-separated file, and df is the DataFrame you will work with.

2. Inspecting Data: Do Not Touch Before You Understand

After loading, your first job is to verify what you actually received. This is where many candidates rush and make wrong assumptions about column names, date formats or missing values.

df.head()
df.shape
df.info()
df.dtypes

3. Selecting Data: Choose the Columns or Positions You Need

Selection answers: “Which columns or cells do I want?” Pandas gives you three common ways.

4. Filtering Data: Keep Only the Rows That Matter

Filtering answers: “Which observations meet my business condition?” The condition returns True or False for every row; Pandas keeps the rows where the condition is True.

premium = df[df["order_value"] > 5000]

south_premium = df[
    (df["region"] == "South") &
    (df["order_value"] > 5000)
]

Worked Example: Filter High-Value Orders

Suppose you load this mini order table.

You want only South region orders above ₹5,000.

result = df[
    (df["region"] == "South") &
    (df["order_value"] > 5000)
]

result[["order_id", "category", "order_value"]]

The row count falls from 5 to 2, which means the filter has done what a business analyst needs - reduced a broad dump into a relevant decision set.

Definitions You Should Be Able to Say Cleanly

PhonePe Pulse: Pandas Basics on Real Indian Payments Data

Situation: India’s UPI ecosystem produces massive transaction activity across states, districts, categories and time periods. PhonePe Pulse made aggregated digital payments patterns easier to explore publicly, giving analysts a rich context for questions such as where adoption is broadening or which transaction categories are gaining visibility.

The move: For a student analyst, the Pandas workflow is straightforward. Load the public dataset files, select the fields that matter - state, district, quarter, transaction count, transaction value and category - and filter for the exact geography or period being studied. This is not a claim about PhonePe’s internal analytics stack; it is the correct external analyst workflow for working with such structured public data.

Outcome or lesson: The primary driver of insight is not the size of the dataset; it is the analyst’s ability to narrow the data to the right business question. Supporting drivers are clean field selection, correct time and geography filters, and validation after every cut. That is exactly what Pandas basics train you to do.

How AI Changes Pandas Basics

AI does not remove the need to know Pandas basics. It changes how quickly you can write, debug and explain them.

Interview Relevance

Common Mistake

What to Revise Next

Once you can load, select and filter, move to the next two layers of the Pandas journey: making the data trustworthy, then summarizing it for decisions.