It’s super simple and clean.īokeh Barchart using Python (via Jupyter Notebook)īefore beginning to use Bokeh, we made the conscious decision to script in JavaScript instead of Python because our entire web application was and is built on a JavaScript framework. Below are examples of barcharts in Bokeh using Python. Experienced Bokeh users can make really beautiful things (see examples here).īut if I wanted to make a visualization that went outside of the conventional characteristics of a chart, such as simulating forces between atoms and dragging the atoms around, I don’t know how I would accomplish that in Bokeh.īokeh is also intended for development in Python, not JavaScript. And that’s okay, because often that’s what the user needs and wants. Bokeh enables you to “chart” in the more conventional sense- it offers a 2-D, grid-like canvas with axes as the baseline. One drawback to Bokeh, however, is that it is limited in the degree of interactivity that a visualization can have. The ability to navigate between Python and JavaScript is also incredibly powerful for charting because Python allows you to tap into useful data structure and data analysis tools, while JavaScript translates the manipulated data into browser-friendly visualizations. It renders first using WebGL with a HTML5 Canvas fallback, provides several built-in tools to interact with charts, handles egregiously large data sets, and ultimately, creates something that can go on the web immediately. In the Bokeh 0.12.13 documentation, it states: “This capability to synchronize between python and the browser is the main purpose of the Bokeh Server.” Source image Advantagesīokeh is magical for a lot of reasons. You can also choose to use the Bokeh Server to handle streaming of your data. Then Bokeh’s JavaScript API takes in your Python script and renders the plots or charts in addition to handling the UI interactions in the browser. You create these visualizations using Python. In an ideal scenario, the library would already include some of these interactive functions that we could easily throw in, instead of having to write them from scratch.īokeh is a library primarily intended for creation of visualizations in the browser from large or streaming datasets.

Our application also needed to have the desired interactivity that we envisioned for the user. Because we were handling larger quantities of data, our visualization had to be able to update at lightning speed (or at least at a speed that had no perceivable lag). Our needs fell into two camps: speed and interactivity. For me, this was “uncharted” territory, and if you feel the same way about charting libraries or data visualization in general, you’ll feel better after reading this. This article will describe our process of choosing a charting library among the myriad of JavaScript charting libraries that are currently available, the critical decision of switching between two charting libraries ( Bokeh and D3.js), and the pros and cons of each one. Perhaps you’re thinking: “What are those options?”, “ How did you approach choosing an option?” or “Why did you feel stupid?” (refers to the Slack message above). More often than not, your initial assumption that a library is the perfect match will be incorrect because of unforeseen obstacles that pop up. However, libraries are never a one-size-fits-all kind of deal. If you weigh your needs correctly and choose a library that somehow satisfies all of them, life is golden. As a result, we had to decide on a library that satisfied our specific use cases. Recently, our team was tasked with creating an interface that needed to integrate a charting library in order to fulfill the goal. When executed correctly, charting libraries have the power to drive home a powerful message and give you the opportunity to view serious visual candy. Several charting libraries are written in JavaScript, a language that is more familiar to web developers than most, which makes learning them a less daunting feat. You are the master of these low level building blocks constituting a “chart”. Think about the charts that you can create within Google Sheets, except now you have direct viewing and editing rights to the library that drives those charts. They are the reason you can quickly grasp trends in life expectancy on FiveThirtyEight or gauge the national sentiment about an upcoming presidential election (yikes) on The New York Times. Navigating *uncharted* watersĬharting libraries produce data-driven visualizations. There comes a time in the life of a budding “low-key but also high-key trying to become a front-end designer and developer” when they must enter the world of charting libraries. By Mandi Cai Charting the waters: between Bokeh and D3 Introduction

Threejs bokeh examples