Technology behind this site
This personal site is built (and now rebuilt) with exclusively open source tools: pelican with a blue penguin theme, a github personal page, and jupyter notebooks. I'm a big fan of not reinventing the wheel, so I spent as much time picking the simplest and generally, most popular tools. I've found this combination maximizes my time sharing content, and minimizes the site maintanence.
A python blogging package. At a high level, this package enables you to write content in markdown files (.rst or .md) and then converts them to html files. These html files are what's pushed online, and serve up static content. It's far simpler to use for a project like this than something designed for enterprise like Django (which I tried first). Best of all, it seems to have gotten more popular over time as evidenced by the number of stars and forks on Github.
There's tons of guides out there already, but I used these two in conjunction with the docs to get me started.
PBPython - An excellent beginner's guide but sacrificing some technical depth for the sake of readability.
Christine Doig's Guide - Most thorough one I've found, but can be a little overwhelming until you familiarize yourself with the package's functionality.
Blue Penguin Theme
A beautiful minimalist theme designed by Jody Frankowski specifically for Pelican. It's barebones and actually removes much of pelican's features (ex: the social media bar) in favor of a very clean, modern look.
Github personal pages
In keeping with the idea to keep it as simple as possible, I choose to make use of Github's personal pages feature . Each account is given one for free, and has a file limit of 1 GB (I'm using about 2% of that). You just need to setup a repo, and you can push directly there. I've configured mine to also redirect to my custom url of alanzzhao.com instead of alanzzhao.github.io.
Even better, pelican has native support for this type of hosting along with Amazon S3 and Dropbox.
Jupyter notebooks are the tool of choice for interactive computing with Python, and increasing for other langauges. They're driven by the powerful idea that data analysis is best served when it comes with insights and the ability for the consumer to experiment further. I wanted to use it as a format to easily share code, visuals, and analysis. Jupyter cuts out the cumbersome tasks of copying and pasting all of those components into a separate markdown file. For the curious, I'll also link the notebooks to be available on Github for reproducibility.
Some blogs write the entire site's content in Jupyter notebooks. Jake Vanderplas's site is an amazing example. From the site format, I strongly suspect he's running pelican as well.
Using Jupyter as a pelican content generation format is easy: it's plugin for pelican . Just install and activate.
Below are some code snippets showing how versatile this platform is (and to prove to myself this worked). Seaborn examples shamelessly taken from the library documentation. Visuals show how three species of flowers are differentiated by characteristics.
import seaborn as sns import pandas as pd %matplotlib inline
sns.set(style="whitegrid", palette="muted") # Load the example iris dataset iris = sns.load_dataset("iris") # "Melt" the dataset to "long-form" or "tidy" representation iris = pd.melt(iris, "species", var_name="measurement") # Draw a categorical scatterplot to show each observation sns.swarmplot(x="measurement", y="value", hue="species", data=iris)
sns.set(style="ticks") df = sns.load_dataset("iris") sns.pairplot(df, hue="species")