Let’s start with the basics.

What is NLP? Natural language processing is the field of using computers to work with text data. It’s what we mean when we talk about AI and human language, and it powers everything from spam filters and chatbots to ChatGPT and voice assistants.

It’s also played a pretty big role in my own career journey.

How It Started

The first time I learned about NLP was back in 2013 during grad school. I had just started my data science career and wanted to work on a portfolio project showcasing my newfound skills. So naturally, I decided to analyze my text messages with my husband… and give it to him as an anniversary gift.

Yes, this really happened, and you can check out the viral blog post here.

I was curious how our conversations had changed over time — from dating to engagement to marriage. I ended up using NLP to look at things like commonly used words, the time of day we sent messages, and more. It was my first real text analysis project, and I fell in love with working with text data.

A Few Years Later…

Fast forward to 2018. I was working at a data science bootcamp and got the opportunity to take a quarter to study something on my own. I chose NLP. A lot had changed since grad school, and I wanted to get up to speed.

That self-study turned into a conference talk on NLP in Python analyzing Ali Wong’s comedy routines, which you can find on YouTube here.

At the time, machine learning was becoming more mainstream, and NLP techniques were evolving quickly — it felt like every few months, a new model was being released. I’d prepare a talk, and then immediately have to update it with the latest method or acronym.

…And Then This Year

In 2025, I was asked to create an NLP course. And I’ll be honest — the learning curve was steep.

The jump from traditional NLP to today’s transformer-based methods was massive. I spent about half a year reading, prototyping, researching, re-reading, and figuring out how to teach it all. What should I include? What should I leave out? How do I break down these concepts in a way that actually makes sense?

It ended up being the most technically challenging course I’ve ever created — but also the one I’m most proud of. You can find the 22-hour-long course here at Maven Analytics, which consolidates over a decade’s worth of NLP knowledge!

The Evolution of NLP

I’ve watched NLP evolve a lot over the years — and the pace of change is only accelerating. Models that were state-of-the-art five years ago feel ancient now.

To make sense of it all, here’s how I organize the three major “eras” of NLP in my mind:

• Early NLP – rules-based methods

• Traditional NLP – text cleaning, machine learning with text classification & topic modeling, etc.

• Modern NLP – transformers, embeddings, attention, LLMs, and Hugging Face

Each era brings its own set of tools and buzzwords, but also its own way of thinking. The main mindset shift I’m seeing is the transition from training your own model to using pretrained models. As a data scientist who practices traditional NLP techniques that use machine learning, it’s essential to understand how the models work so I can interpret them correctly.

Now, it’s a completely different ballgame with modern NLP.

We’ve gotten to the point where LLMs are so complex that it’s impossible to understand what’s happening under the hood — and that’s okay. As a data scientist, the norm is to now use pretrained models (even if you don’t completely understand how they work), incorporate them into your analysis, and potentially fine-tune them to make the results make more sense.

In my opinion, this is one of the biggest shifts in data science over the past decade.

What’s in the NLP in Python Course?

If you’re interested in diving deeper into NLP, here’s what I cover in my course:

• NLP 101 – how the field has changed, what’s worth learning, and what you can probably skip

• Text Preprocessing – all the ways to clean and prepare your data

• Machine Learning – how to build predictive models and do things like topic modeling

• Neural Networks & Deep Learning – a step-by-step walkthrough of how they work

• Transformers – deep dive into embeddings, attention, and feedforward networks

Hugging Face – how to use modern pretrained LLMs in Python

My goal was to make the hard stuff feel a little less intimidating — to give you not just the “how,” but also the “why.” I want you to understand what’s going on under the hood so you can build intuition that lasts beyond the next trending model.

What’s Next for NLP

Honestly? No one really knows, but it’s probably headed toward more multimodal, assistant-style tools that feel collaborative and intuitive. NLP is moving so fast, and we’re all just trying to keep up these days.

If you’re feeling overwhelmed — that’s normal. I’ve been in this space for over a decade, and I still feel like I’m constantly learning. What’s helped me is focusing less on memorizing every new tool, and more on understanding the foundations. That’s what I’ve tried to build into my NLP in Python course.

I hope it helps you navigate the wild, wonderful world of NLP — whether you're analyzing your own text messages or building the next viral chatbot.

And if you do check out the course — I’d love to hear what you think!