I didn’t start my career with Python or fancy neural networks. Nope, I was deep in the world of spreadsheets, financial models, and Excel shortcuts – until the world of data science called my name. If you’re in a similar boat – coming from finance or another field and learning data science on the job – this post is for you. Here’s how I went from financial reports to building machine learning models… and how you can do it too.

1. The Foundation: How I Went from Excel Wizard to Data Science Beginner

My first steps into data science felt like trying to walk before I could crawl. I knew I had to level up beyond pivot tables.

Core Skills to Master: Learning Python was key. If you can’t wrangle data with pandas and NumPy, forget about deep learning models. SQL was also critical – master it early.

Key Concepts: Statistics didn’t click immediately, but eventually, it became essential. Concepts like regression felt familiar, but clustering was a tough one to crack.

Best Resources: Online courses were crucial. StackOverflow helped me troubleshoot, GitHub connected me with real projects, and ChatGPT became my 24/7 “friend”.

2. Building a Portfolio (That Doesn’t Involve Spreadsheets)

I was used to financial forecasts in Excel, but for data science, I had to raise the bar.

Importance of Projects: I realized projects were the best way to prove my skills. I started analyzing real-world data and building models – though predicting stock prices taught me a tough lesson.

Making Your Portfolio Stand Out: Document your process. Don’t just upload code, explain your reasoning and how you approached problems.

3. Navigating the Tools of the Trade (AKA Ditching Excel)

It wasn’t just about Python – I had to master the tools data scientists use.

Software and Libraries to Know: TensorFlow, PyTorch, scikit-learn, pandas, Matplotlib, Seaborn – these became my toolkit. They replaced the old Excel ways.

Automation and Workflow: I’ve moved past manual Excel updates. Automating tasks using Microsoft tools and exploring options like Airflow and Docker are on my radar.

4. Advanced Topics (Where Things Get Really Interesting)

With a solid foundation, it was time to dive deeper into advanced topics.

Deep Learning: Neural networks offer great potential for pattern recognition. Always start with clean data, especially for pre-processing and feature scaling.

Natural Language Processing (NLP): NLP is crucial for analysing unstructured text. Pre-trained models like BERT or GPT save time and give strong results when fine-tuned.

AI Ethics: Bias in models can skew decisions. Using SHAP and fairness metrics helps ensure your models are fair and reliable.

5. Networking and Building My Brand (Because You Can’t Do This Alone)

Learning is important, but connecting with others is just as essential.

Building Relationships: Networking through LinkedIn, meetups, and open-source contributions made a big difference. Feedback from the community was invaluable.

Job Search Tips: Show not just your code, but your problem-solving process. Hiring managers value critical thinking.

Personal Branding: I started sharing my work online – whether through blog posts, LinkedIn, or meetups. It’s about showcasing your skills and putting yourself out there.

So, here I am – half-finance, half-data science, all-in on learning. The journey wasn’t always easy, but it was worth every line of code and every late-night debugging session. If you’re in a similar situation, just know: the skills you already have in finance can give you a huge head start in data science. All you need is the courage to dive in and start learning.
And don’t worry, spreadsheets will always have a special place in my heart. But now, I’m all about that Python-powered future!

Want to Apply This in Your Work?

If you’re working with data and want to go beyond basic charts and scripts, I offer tailored support to help you build real impact with your insights:

🎓 Personalized Workshops: Hands-on training sessions tailored to your team, tools, and goals

🧠 Consulting Services: Strategic guidance on turning data into decisions – whether it’s modelling, metrics, or storytelling

🕒 Fractional Data Scientist: Embedded support for companies who need senior data expertise without a full-time hire

Whether you’re scaling a product team or building out your first data processes, I can help you use data more effectively, with clarity and confidence.

👉 Get in touch to start a conversation.