Data Visualization

Data visualizations transform raw numbers into graphic formats that make it easier for humans to see patterns, trends and other useful information.

Python data visualization tools

• Bokeh

• HoloViews

• Matplotlib

• Chartify (source code)

• Graphviz

Python-specific data viz resources

• Python Data Visualization 2018: Why So Many Libraries? is an in-depth article on the Python data visualization tools landscape. A must-read whether you are new to the space or have been using one or more of these libraries for awhile.

• The Python Graph Gallery has a slew of visualizations created with Python and includes the code used to produced each one.

• Python & OpenGL for Scientific Visualization is a free book that shows how to combine open source tools such as PyOpenGL with Python data analysis libraries to generate interactive scientific data visualizations.

• 10 Useful Python Data Visualization Libraries for Any Discipline is a straightforward overview of Python packages that create Python visualizations.

• Introduction to Data Visualization with Altair is a starter post for the wonderful Altair visualization tool written in Python.

• The Next Level of Data Visualization in Python uses the Plotly graphing library to draw more complex visualizations.

• An introduction to Altair provides another wonderful tutorial on this data visualization tool.

• A Dramatic Tour through Python’s Data Visualization Landscape provides examples with the ggplot and Altair libraries. The question-and-answer format for what you can do with the data is a really good model that keeps your attention throughout the post.

• How to Generate FiveThirtyEight Graphs in Python gives a great tutorial on generating a specific style graph with pandas and Matplotlib that is similar to FiveThirtyEight's plots.

• Intro to pdvega - Plotting for Pandas using Vega-Lite shows how to generate plots from your pandas-structured data using pdvega.

• Sorting Algorithms Visualized in Python uses Python, numpy and scikit-image to animate how sorting algorithms work.

• How to Build a Reporting Dashboard using Dash and Plotly explains how to use the Dash library to take a bunch of data and turn it into a nice-looking dashboard.

• The reason I am using Altair for most of my visualization in Python explains why this wrapper for Vega-lite is awesome and that the author uses Altair because Matplotlib can be very complicated for whipping up quick visualizations.

Beautiful example visualizations

Sometimes you need inspiration from other sources to figure out what you want to build. The following links have made me excited about data visualization and gave me ideas for what to build.

• Monarchs is a wonderful 1,000 year history visual of European rulers. The developer also wrote an in-depth article on how Monarchs was created using d3.js.

• Star Wars: The Force Accounted is Bloomberg's way of breaking down on-screen action between light and dark sides, the main characters, various bits about the Force and other data extracted from the movies.

• What do numbers look like? is a Python 3 dimensional visualization of millions of integers, colored by special factors such as prime and Fibonacci numbers.

• Bay Area Housing Marketing Analysis: Part 1 and Part 2 are a combination of inspiration and tutorial. These posts contain a ton of data analysis and graphing and show numerous ways to slice and present information.

• How We Animated Trillions of Tons of Flowing Ice breaks down the process that the NY Times data team used to create the beautiful Antarctic Dispatches articles that show how glaciers and ice are moving.

• Who knew good old histograms could be so fascinating? Check out this post titled What's so hard about histograms? and scroll through to learn a ton about the details you can think about when creating these types of data visuals.

• Optimized Brewery Road Trip, With Genetic Algorithm shows how a heuristic solving approach such as a genetic algorithm can be used to handle a version of the Traveling Salesman Problem (TSP), but with the more fun Top 100 American brewery locations.

Data visualization resources

• Guides for visualizing data provides the thought processes that you can use when you are showing uncertainty, incompletet data, differences, outliers, and other common scenarios that occur when your data is messy - and it usually is!

• Data visualization, from 1987 to today is a wonderful reference about the pre-computer age era of visualization which was a combination of cartography, art and statistics rather than any cohesive field as it is often seen today. The images showing how people worked with paper to build their visuals add fantastic context to the story.

• Xenographics presents uncommon and unusual visualization formats such as the Manhattan Plot and Time Curve.

• Engineering Intelligence Through Data Visualization at Uber explains how Uber's data visualization team grew from 1 person to 15 and the output they created along the way, including the open source tool react-vis.

• The Practitioner's Guide to System Dashboard Design series covers a lot of ground for what you should consider when building one form of visualization, the data dashboard:

  • Part 1: Structure and Layout
  • Part 2: Presentation and Accessibility
  • Part 3: What Charts to Use
  • Part 4: Context Improvement

• Truncating the Y-Axis: Threat or Menace? is a great in-indepth explanation of why something that seems as straightforward as laying out the Y-Axis requires thought and care otherwise the visualization could end up being misleading.