Teach Yourself Data Science: The Learning Path I Used to Land an Analytics Job at Jet.com

So you want to learn data science and break into the tech industry, but you don‘t know where to start? I was in the same position a few years ago. As a finance major, I never thought I‘d be working in analytics at an ecommerce company, using SQL and Python to wrangle data every day. But through self-teaching and a lot of hard work, I made it happen. Here‘s the learning path I followed to gain the skills to land a job as a data analyst at Jet.com.

Cultivating the Right Mindset

Before diving into the technical skills, it‘s crucial to develop the right mindset and motivation. Teaching yourself data science is a long journey that will inevitably have many challenges and moments of frustration. You need to be prepared to push through these obstacles.

For me, setting aside dedicated time to learn, around 1-2 hours per day, helped build consistency. I also found it helpful to remind myself of my end goal – to build the skills to transition into a data-focused role. Keeping sight of this motivation pushed me to keep going even when the concepts seemed too difficult.

It‘s also important to understand that you won‘t become an expert overnight. Focus on consistent incremental progress and celebrate small wins along the way. With the right mindset and discipline, you‘ll be amazed at how much you can teach yourself over several months of focused learning.

Core Skills Overview

So what exactly do you need to learn to be a data scientist? At a high level, the core skills are:

• Programming, especially Python and/or R
• Databases and SQL
• Statistics and probability
• Machine learning
• Domain expertise (varies by industry)

Below I‘ll dive into each and share the best resources I found for self-teaching.

Python Programming

For data science, Python is the most popular and versatile programming language to learn. It has powerful libraries for data analysis, visualization, machine learning, and more. And compared to other languages, it‘s more intuitive to pick up.

The best free introduction to Python in my opinion is Zed Shaw‘s Learn Python the Hard Way course. It‘s entirely self-directed and the material makes programming feel approachable, giving you confidence that you can do this. Once you have the basics down, you can start doing data analysis with Python through libraries like Pandas. The Mode Analytics Python tutorial is a great next step after Learn Python the Hard Way.

SQL and Databases

Most data at companies lives in databases like MySQL and Postgres. To be able to explore and analyze this data, you need to be proficient in SQL – the language used to query databases.

Mode Analytics has an excellent free SQL tutorial that covers all the key concepts. They even provide a SQL editor and example database to practice with. I also frequently referenced W3 School‘s SQL material when I got stuck. With some practice, you‘ll be able to extract data, join tables, and conduct complex analysis all through SQL.

Machine Learning

Machine learning is an exciting and powerful field that uses algorithms to make predictions, uncover insights, and build intelligent applications. To get started with ML, I recommend the free Udacity Intro to Machine Learning course. It provides a solid foundation on core concepts and the most commonly used algorithms.

From there, you can go deeper with resources like Stanford‘s CS231n on convolutional neural networks (used in computer vision) and the book Grokking Deep Learning for clear explanations of neural networks. It‘s also worthwhile to learn deep learning libraries like TensorFlow, Google‘s popular open-source library for machine learning. Their MNIST tutorial is a great intro.

Statistics

Having a strong grasp on statistics and probability is crucial for making rigorous conclusions from data. Key concepts to learn include statistical significance, distributions, hypothesis testing, regression analysis, and Bayesian methods.

To learn these, I took the Udacity Intro to Statistics course. For more depth, an excellent book is Think Stats, which explains concepts through practical examples in Python. You don‘t need a PhD in statistics to be a data scientist, but the stronger your stats foundation, the better you‘ll be at reasoning with data.

Tying It Together: Building Projects

It‘s one thing to take courses, but the best way to solidify and demonstrate your data science skills is through projects. As you‘re learning, work on data projects that interest you. Find a public dataset, formulate an interesting question or problem, and use your programming, analysis, and modeling skills to find insights.

For example, you could build a classifier to categorize emails as spam/not spam, analyze NYC subway traffic data to optimize commute times, or build a Twitter sentiment analysis tool. The possibilities are endless – the key is to pick something you‘re curious about so you‘ll stick with it. Document your project and thought process in Jupyter Notebooks or Github. This builds your data science portfolio to show potential employers.

Another great way to practice is through Kaggle, a platform for data science competitions. Companies post data challenges and users submit their best models for prizes and bragging rights. Kaggle has tons of interesting datasets and kernels shared by other data scientists – it‘s a fantastic way to learn and test your skills.

Structured Programs

For those who prefer more structured curriculums, several MOOCs can help accelerate your learning path:

• Udacity Data Analyst Nanodegree
• Udacity Machine Learning Engineer Nanodegree
• Coursera Applied Data Science with Python Specialization
• DataCamp Data Scientist Track

These are paid programs, but they provide well-crafted curriculums, hands-on projects, and community support. I took Udacity‘s Data Analyst program and found it very beneficial for guided project-based learning. If you have the budget, these programs can add good structure and accountability to your self-learning journey.

Immerse Yourself

Becoming a data scientist isn‘t just about technical skills – it helps tremendously to immerse yourself in the data science community. Find data science blogs, podcasts, and newsletters to follow and read/listen to consistently. Some of my favorites:

• Data Skeptic podcast
• Data Science Weekly newsletter
• Kaggle blog
• KDnuggets blog

It‘s also hugely beneficial to meet others on similar learning journeys or already working in data science. Attend data science meetups in your area to learn from talks and network. Participate in local data science organizations. The more you engage with the community, the more you‘ll learn from others further along the path and stay motivated.

Bringing It to the Job

Once you‘ve built your foundational data science skills, start looking for ways to apply them on the job. Perhaps there are datasets sitting around that you could dig into and uncover insights to help your team. Automate analysis and data tasks using your programming skills. Incorporate machine learning into a company process. Volunteering for data projects, even as a side project at first, shows initiative and gives you valuable real-world experience.

The data science skills you‘ve developed are extremely transferable and valuable across industries – you just have to seek out ways to apply them within your domain. Over time, you can take on more data-focused responsibilities and make the transition into a full-time data scientist role, whether at your current company or a new one. This was my path – starting with ad-hoc data analysis and eventually moving into a dedicated analytics role.

Never Stop Learning

Data science is a vast and rapidly evolving field. The learning doesn‘t stop once you land the job. Commit to continuously developing your skills and knowledge. Stay on top of industry trends. Pick up new programming languages and tools as they emerge.

The beauty of data science is that there is always more to learn. While the self-teaching journey is challenging, it‘s also deeply rewarding as you bring together technical skills to solve real-world problems and make data-driven decisions. Through hard work, discipline, and a genuine curiosity, you can absolutely teach yourself data science and open up a world of career possibilities. So get started and enjoy the ride!