Python for Data Vis

Steven Bergner

September 17, 2018

Goal

• Explore data processing options in Python
• Compare a few visualization modules
  • Jupyter Notebook / Lab
  • Matplotlib
  • IPyWidgets
  • Plotly / Dash

Programming in Data Science

• Hadley Wickham: “Can’t do data science in a GUI”
• Programming languages are Languages (%>%)
• Stored as Text
• Copy & Paste
• Provenance: Reproducible, Readable, Diffable, Open

Why Python?

• World’s most popular coding language
• HN Language comparison

Python Vis Landscape

• Jake VanderPlas’ presentation

Jupyter

• Consider Jupyter Notebook and Lab

  jupyter lab --no-browser --port=8889
  

IPyWidgets / Jupyter Widgets

• Jupyter Dashboard Layouts
• IPyWidgets Documentation

QGrid

• View, Filter, and Edit large DataFrames
• Quantopian/QGrid
• Introduction
• Other Quantopian Intros

Data for Experimentation

• Global Landslide Catalog Export
• See Kirschbaum et al. for analysis example:
  • Landslide Catalog: Method and Results
  • Presentation on susceptible areas
• Combine with elevation information (code)

Visualize This! Challenge

• Westgrid Visualization Challenge
• Register by Oct 1^st to participate