Google Colab

Notebooks Explore and run machine learning code with Kaggle Notebooks, a cloud computational environment that enables reproducible and collaborative analysis Types of Notebooks There are multiple different types of Notebooks on Kaggle. Notebooks The most common type is Jupyter notebooks (usually just “notebooks”). Jupyter notebooks consist of a sequence of cells, where each cell is formatted in either Markdown (for writing text) or in a programming language of your choice (for writing code). To start a notebook, click on Create > New Notebook. This will open the Notebooks editing interface. Notebooks may be written in either R or Python. The language can be changed under the File > Language menu. Scripts It's also possible to create a script. Scripts are files that execute everything as code sequentially. To start a script, create a new notebook, and then change the type to Script under File > Editor Type. Also under the File menu, you can select what type of script you would like to execute. You may write scripts in R or in Python. You can also execute selected lines of code by highlighting the code in the editor interface and clicking the “Run” button or hitting shift-enter. Any results will be printed to the console. RMarkdown Scripts RMarkdown scripts are a special type of script that executes not just R code, but RMarkdown code. This is a combination of R code and Markdown editing syntax that is prefered by many R authors in our community. The RMarkdown editor is the same one used for basic R or Python scripts, except that it uses the special RMarkdown syntax. To start editing an RMarkdown script, create a new notebook, change the file type to "Script" under File > Editor Type and then change the language under File > Language Searching for Notebooks In addition to being an interactive editing platform, you can find and use code that others in the community have shared public. Kagglers working with data across both the Datasets and Competitions platforms are constantly building cool things. Exploring and reading other Kagglers’ code is a great way to both learn new techniques and stay involved in the community. There’s no better place than Kaggle Notebooks to discover such a huge repository of public, open-sourced, and reproducible code for data science and machine learning. The latest and greatest from Notebooks is surfaced on Kaggle in several different places. Site Search You can use the site search in the top bar of the website while on any page to look for not only Notebooks but Datasets, Competitions, Users, and more across Kaggle. Start typing a search query to get quick results and hit "Enter" to see a full page of results that you can drill down into. From the full page search results, you can filter just to "Notebooks" and add even more filter criteria using the filter options on the left hand side of the page. Homepage When you’re logged into your Kaggle account, the Kaggle homepage provides a live newsfeed of what people are doing on the platform. While Discussion forum posts and new Datasets make up some of the contents of the home page, most of it is dedicated to hot new Notebooks activity. By browsing down the page you can check out all the latest updates from your fellow Kagglers. You can tweak your newsfeed to your liking by following other Kagglers. To follow someone, go to their profile page and click on “Follow User”. Content posted and upvotes made by users you have followed will show up more prominently. The same is true of other users who choose to follow you. Post high-quality notebooks and datasets and you will soon find other users following along with what you are doing! Notebook Listing A more structured way of accessing Notebooks is the Notebook listing, accessible from the “Notebooks” tab in the main menu bar. The Notebook listing is sorted by “Hotness” by default. “Hotness” is what it sounds like: a way of measuring the interestingness of Notebooks on the platform. Notebooks which score highly in Hotness, and thus appear highly in this list, are usually either recently written Notebooks that are scoring highly in things like upvotes and views, or “all-time” greats that have been consistently popular on the platform for a long time. Other methods of sorting are by Most Votes: Surfaces the most popular notebooks of all time Most Comments: Returns the most discussed notebooks of all time Recently Created: A real-time stream of new Notebooks Recently Run: A real-time stream of activity Relevance: Sorts the results based on their relevance to the query Other filtering options, available from the navigation bar, are Categories (Datasets or Competitions?), Outputs, Languages (R or Python?), and Types (Script or Notebook?). You can also use the Notebook listing to sort through your own Notebooks (“Your Work”), find Notebooks that others have shared with you ("Shared With You"), or to look at Notebooks you have previously upvoted (“Favorites”). Finally