Matplotlib: Difference between revisions

From CommunityData
 
 
(15 intermediate revisions by 5 users not shown)
Line 1: Line 1:
[[File:Matplotlib-hist2d.png|right|500px]]
[[File:Matplotlib-hist2d.png|right|500px]]
__NOTOC__
__NOTOC__
== Visualizing data with Matplotlib and Wiki-bios ==
== Getting Started ==
Download the CDSW Matplotlib code from [https://github.com/makoshark/matplotlib-cdsw/archive/master.zip here] and unzip it on your machine.
 
We will also need the data file that is [http://communitydata.cc/~mako/hp_wiki.tsv here]
 
== Visualizing data with Matplotlib ==


In this session, we will explore how to produce clear, informative charts, graphs, and plots with [http://matplotlib.org/ Matplotlib], the most popular toolkit for scientific data visualization in Python.
In this session, we will explore how to produce clear, informative charts, graphs, and plots with [http://matplotlib.org/ Matplotlib], the most popular toolkit for scientific data visualization in Python.


We'll be focusing on a dataset drawn from Wikipedia and , containing the names, birth dates, genders, article creation dates, and number of edits, of over 180,000 Wikipedia biography articles.
We'll start with the data-set created [[Community_Data_Science_Workshops_(Fall_2015)/Day_3_Lecture|this morning]] containing information about edits to the Harry Potter Wikipedia article.  


The dataset used here has been drawn from several sources. The list of biographies and the gender of the biography subjects comes from a datbase called [http://dbpedia.org DBpedia] that is a structured database built from Wikipedia. If determines if articles are biographies from the Wikipedia category system and it determines if the subjects are male or female primarily by looking at whether subject is referred to as ''he'' or ''she'' in the article. Birthdates for people are extracted from the data recorded in Wikipedia in the "infobox" sidebars common to many pages. Data on how people edit Wikipedia is merged onto this dataset from a separate collection of metadata of Wikipedia editing very similar to the one we built for Harry Potter articles the morning.
We will then proceed to visualize different aspects of data from the [[Community_Data_Science_Workshops_(Fall_2015)/Day_2_Projects/Socrata|Socrata web API]].
 
 
=== Inspiration ===
 
* [https://flowingdata.com/ Flowing Data]
* [http://www-01.ibm.com/software/analytics/many-eyes/ Many Eyes]
* [http://www.edwardtufte.com/tufte/ Edward Tufte]
* [http://www.visualizing.org/ visualizing.org]
* [http://idl.cs.washington.edu/ UW Interactive Data Lab]


=== Goals ===
=== Goals ===
Line 19: Line 33:
* Exercise your creativity by making your own visualization
* Exercise your creativity by making your own visualization


=== Download and test the Matplotlib-with-Wiki-bios project ===
# First plot: <tt>001-hello-plot.py</tt>
 
# Subplots: <tt>002-subplots.py</tt>
(Estimated time: 10 minutes)
# Let's do something more interesting: <tt>003-plot-timeseries.py</tt>
 
# Visit the [http://matplotlib.org/gallery.html Matplotlib gallery].
After installing matplotlib, and downloading and unpacking the Wikibios bundle, move into that directory with '''cd'''.  You can test your installation by running '''python histograms.py'''.  If matplotlib is install correcting, a chart file named '''histograms.pdf''' will appear in the current directory.
# Make another kind of plot: <tt>004-plot-histogram.py</tt>
 
# Dive deeper into web APIs: <tt>005-traffic-timeseries.py</tt>
Bundle with code for all platforms: http://mako.cc/teaching/2014/cdsw-autumn/wikibios.zip
# Play around with any/all the data you've seen! You can find some more examples in the <tt>wikibios</tt> folder.
 
=== References ===
=== References ===


Line 32: Line 46:
* [http://matplotlib.org/examples/index.html matplotlib Examples] (many, with source)
* [http://matplotlib.org/examples/index.html matplotlib Examples] (many, with source)
* Other plotting resources
* Other plotting resources
** [http://blog.olgabotvinnik.com/prettyplotlib/ prettyplotlib]: hip-aesthetic matplotlib plots
** [http://web.stanford.edu/~mwaskom/software/seaborn/ Seaborn]: fancy matplotlib-based visualizations
** [http://web.stanford.edu/~mwaskom/software/seaborn/ Seaborn]: fancy matplotlib-based visualizations
** [http://ggplot.yhathq.com/ ggplot]: port of the R language's ggplot2 library to python
** [http://ggplot.yhathq.com/ ggplot]: port of the R language's ggplot2 library to python
Line 48: Line 61:


[[File:Wikipedia.png|right|250px]]
[[File:Wikipedia.png|right|250px]]
[[Category:CDSW]]

Latest revision as of 16:22, 7 May 2016

Getting Started[edit]

Download the CDSW Matplotlib code from here and unzip it on your machine.

We will also need the data file that is here

Visualizing data with Matplotlib[edit]

In this session, we will explore how to produce clear, informative charts, graphs, and plots with Matplotlib, the most popular toolkit for scientific data visualization in Python.

We'll start with the data-set created this morning containing information about edits to the Harry Potter Wikipedia article.

We will then proceed to visualize different aspects of data from the Socrata web API.


Inspiration[edit]

Goals[edit]

  • Get set up to make graphs with Matplotlib
  • Learn the basics of the Matplotlib API and workflow
  • Practice reading the Matplotlib documentation
  • Build a plotting program step by step
  • Learn simple ways to distill the essence of a large data set
  • Explore the art of visualizing data
  • Exercise your creativity by making your own visualization
  1. First plot: 001-hello-plot.py
  2. Subplots: 002-subplots.py
  3. Let's do something more interesting: 003-plot-timeseries.py
  4. Visit the Matplotlib gallery.
  5. Make another kind of plot: 004-plot-histogram.py
  6. Dive deeper into web APIs: 005-traffic-timeseries.py
  7. Play around with any/all the data you've seen! You can find some more examples in the wikibios folder.

References[edit]

Example topics to cover in Lecture[edit]

  • line charts
  • histograms
  • binning
  • scatter plots
  • heat maps
  • axis labeling
  • legends