Practical Machine Learning with Kaggle
I always liked to learn anything in action. It feels so great to see the results immediately. Also, I can gain a better understanding of the problem. I did the same for machine learning. I’ve been developing applications for years and I wanted to jump into the Machine Learning world. There were a lot of theories that I need to learn for sure, but I needed something to spice up my learning and get some excitement. So, I ended up Kaggle.
What is Kaggle?
First, let me ask you some questions.
- When you start learning, you will need databases. So that you can explore data and find relations and features of data. So, if you don’t have data, you probably won’t learn anything. Isn’t it bad?
- Consider you have the data. Now, you might end up with a bunch of questions and you might don’t know what to do next? How you can deal with this data and export some meaningful results? You might need to ask a friend who is more experienced. Where can you find this friend?
- Ok, you have data, you have a friend to ask, but how you can find a real-life problem to solve using your machine learning techniques?
- You have data, an experienced friend, and a real-life problem. Can you run your analysis on your computer? Is it powerful enough? How do you want to deal with big amounts of data? Do you really think your fancy laptop can handle it? Yes, it probably can. But, you might need a new laptop sooner. Because of overheating the CPU!
There is more to ask, but there’s one answer. Kaggle. Here are some features of Kaggle that you might find exciting when you are learning Data Science.
- Over 50,000 public datasets
- A large community of data scientists
- Public code samples
- Online python code editor
- Access to a cloud-based GPU and CPU that are way better than our laptops for sure
- Free online micro-courses for Python, Pandas, Machine Learning, AI, and more.
- Job postings for Data Science positions
- Online Data Science competitions with prizes
Take a look at this video by Kaggle. You might get a better idea: