Introduction to Kaggle Kernels

YUFENG GUO: On this episode of AI Adventures, find out what Kaggle Kernels are and how to get started using them. Though there's no popcorn in this episode, I can assure you that Kaggle Kernels are popping. Kaggle is a platform for doing and sharing data science. You may have heard about some of their competitions, which often have cash prizes. It's also a great place to practice data science and learn from the community. Kaggle Kernels are essentially Jupyter Notebooks in the browser that can be run right before your eyes, all free of charge. Let me say that again in case you missed it, because this is truly quite amazing. Kaggle Kernels is a free platform to run Jupyter Notebooks in your browser. This means that you can save yourself the hassle of setting up a local environment and have a Jupyter Notebook environment right inside your browser anywhere in the world that you have an internet connection. Not only that-- the processing power for the notebook comes from servers up in the clouds, not your local machine. So you can do a lot of data science and machine learning without heating up your laptop. Kaggle also recently upgraded all their kernels to have more compute power and more memory, as well as extending the length of time that you can run a notebook cell to up to 60 minutes. But OK. Enough of me gushing about Kaggle Kernels. Let's see what it actually looks like. Once we create an account at Kaggle.com, we can choose a dataset that we want to play with and spin up a new kernel or notebook in just a few clicks. The dataset that we started in comes preloaded in the environment of that kernel, so there's no need to deal with pushing a dataset into that machine or waiting for large datasets to copy over a network. Of course, you can still load additional files into the kernel if you want. In our case, we'll continue to play with our fashion and this dataset. It's a dataset that contains 10 categories of clothing and accessory types-- things like pants, bags, heels, shirts, and so on. There are 50,000 training samples and 10,000 evaluation samples. Let's explore the dataset in our Kaggle Kernel. Looking at the dataset, it's provided on Kaggle in the form of CSV files. The original data was in a 28 by 28 pixel grayscale images and they've been flattened to become 784 distinct columns in the CSV file. The file also contains a column representing the index, 0 through 9, of that fashion item. Since the dataset is already in the environment, in pandas-- this is already loaded-- let's use it to read these CSV files into panda's data frames. Now that we've loaded the data into a data frame, we can take advantage of all the features that this brings, which we covered in the previous episode. We'll display the first five rows with Head, and we can run Describe to learn more about the structure of the dataset. Additionally, it would be good to visualize some of these images so that they can have more meaning to us than just rows upon rows of numbers. Let's use matplotlib to see what some of these images look like. Here we'll use the matplotlib.pyplot library-- typically imported as PLT-- to display the arrays of pixel values as images. We can see that these images, while fuzzy, are indeed still recognizable as the clothing and accessory items that they claim to be. I really like that Kaggle Kernels lets me visualize my data in addition to just processing it. So Kaggle Kernels allows us to work in a fully interactive notebook environment in the browser with little to no setup. And I really want to emphasize that we didn't have to do any sort of Python environment configuration or installation of libraries, which is really cool. Thanks for watching this episode of Cloud AI Adventures. Be sure to subscribe to the channel to catch future episodes as they come out. Now what are you waiting for? Head on over to Kaggle.com and sign up for an account to play with kernels today. [BEEP] Though there's no popcorn in this episode, I can assure you that Kaggle Kernels-- [BEEP] You've got to throw harder. SPEAKER: That's horrible timing. [BEEP] YUFENG GUO: Wait, are you going to throw it this way or this way? [BEEP] Though there's no popcorn in this episode, I can assure you that [LAUGHING] Kaggle Kernels are popping.

Channel: Google Cloud Tech

Views: 92,299

Rating: 4.9439654 out of 5

Keywords: kaggle kernels, kaggle, data science, tensor flow, machine learning, AI, jupyter notebooks, google cloud, cloud platform, cloud developers, jupyter, artificial intelligence, mnist, fashion mnist, web developer, data science results, python, kernel, big data ai, ai adventures, data analysis, cloud data, TensorFlow, big data, gcp machine learning, training, estimators, classification, linear classification, machine learning models, fullname: Yufeng Guo, GDS: Yes;

Id: FloMHMOU5Bs

Length: 4min 21sec (261 seconds)

Published: Wed Dec 06 2017