Pandas Data Analysis

This week, I dove deep into the world of data analysis using Pandas — one of the most powerful and beginner-friendly Python libraries for working with structured data. I didn’t aim to master the syntax or teach anyone else, but instead to understand how real-world data "feels" when you work with it, and to develop some raw insights through exploration.

---

What I Built

I worked on three real-world projects, each using public Kaggle datasets:

1) IMDb Dataset Analysis

→ Extracted top-rated movies, genre trends, decade-wise analysis, and director/actor breakdowns.

2) Netflix Dataset Analysis

• Explored global content distribution, genre frequency, release timelines, and show vs. movie split.

• https://github.com/Kartikshard/Pandas-analysis/blob/main/Netflx_Analysis.ipynb

3) Steam Games Insight

Analyzed user ratings, game popularity, release patterns, and pricing strategies in the gaming industry.

Resources I Used

• IMDB Dataset of 50K Movie Reviews

• Steam Games Dataset

• Netflix Movies and TV Shows Dataset

• Stack Overflow (a lot!)

• Official Pandas documentation

---

Problems I Faced

• Data Cleaning was messy. Each dataset had quirks — missing values, mixed data types, duplicate rows — that slowed me down at first.

• Column naming conventions differed, making it tough to generalize across projects.

• Struggled to visualize relationships early on before learning how to properly structure .groupby() and pivot tables.

• Analysis paralysis: I initially tried to answer too many questions. Narrowing my focus made things much clearer.

---

Key Observations

• Pandas is extremely intuitive once you stop thinking like a coder and start thinking like a data analyst.

• Small data decisions (like how to handle nulls) shape the entire outcome of a project.

• Visualization and storytelling should start earlier in the project, not just after analysis is complete.

• Real-world datasets are rarely clean. Handling that mess is the skill.

---