Getting Started with Python

Exploratory Data Analysis

Alp Tezbasaran

2026-02-10

Welcome

Slides: https://go.ncsu.edu/dss-python QR code to workshop material

To get a complete from this session:

In-person: Sign the attendance sheet
Virtual: We will use Zoom attendance

Who we are

Mara Blake

Head, Data Science Services

Shannon Ricci

Associate Head, Data Science Services

Jeff Essic

Lead Librarian for GIS

Alp Tezbasaran

Data Science Librarian

Selene Schmittling

Data Science Librarian

Lynnee Argabright

Data Science Teaching Librarian

Our team comes from a variety of academic backgrounds, including nuclear engineering, electrical engineering, marine ecology, history, information science, natural resources, and geographic information sciences.

Check our workshops

go.ncsu.edu/datavizworkshops QR code to Data Viz Workshops

Mapping and Analyzing Urban Heat Island Effects with ArcGIS Online
Designing Data Visualizations for Research Presentations and Publications
Getting Started with Python [R]: Fundamentals
Getting Started with Python [R]: Loops and Functions
Exploratory Data Analysis with Python [R]
Git Moving: A Beginner’s Guide to Version Control
Colab and Chill: A Vibe Coding Approach to Exploratory Data Analysis using AI

Have workshop ideas? Contact us!

How to contact us

go.ncsu.edu/getdatahelp

Get Data Help

Direct email: getdatahelp@ncsu.edu

Who is this workshop for?

Learners who know basic Python syntax (variables, lists, simple conditionals/loops)
Anyone who wants to build confidence exploring tabular data in a repeatable way
People who use spreadsheets and want a Python workflow for analysis

Best fit in the full Getting Started with Python sequence:
- Getting Started with Python - Fundamentals
- Getting Started with Python - Loops and Functions
- Exploratory Data Analysis with Python

What is exploratory data analysis?

Goal: understand quality, structure, and patterns before modeling
Typical cycle: inspect -> clean -> visualize -> summarize -> refine
Outcome: better questions and better next analysis steps

EDA Concepts

Shape: rows and columns
Schema: column data types
Missingness: null counts and percentages
Distribution: center, spread, skew, outliers
Cardinality: number of unique values
Grain: what one row represents
Association: variables that move together (not causation)

What can EDA help us decide?

Are we ready to analyze, or do we need more cleaning?
Which variables look useful, redundant, or suspicious?
Which transformations/features should we try next?
Which subgroups need deeper analysis?
What should come next: statistical test, model, or new data collection?

Important boundary

EDA helps generate evidence and hypotheses, but it does not prove causation.

Why Polars in this workshop?

Fast and memory-efficient for tabular workflows
Expression-based API keeps transformations readable and teachable
Great for core data-frame verbs (select, filter, with_columns, group_by)
Works well with familiar plotting tools like matplotlib

Polars vs pandas (quick view)

Topic	Polars	pandas
Execution style	Expression-oriented, optional lazy mode	Eager by default
Performance	Often faster on larger tabular transforms	Strong general-purpose baseline
Teaching focus today	Clear pipeline thinking	Familiar ecosystem many already know
Recommendation	Use whichever fits your workflow/team	Concepts transfer between both