Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Telling stories with data - PyCon Sweden 2017

Eleonore
February 08, 2024

Telling stories with data - PyCon Sweden 2017

Use the Altair library for a fast data exploration and visualization in Python and a shareable web visualization output.

Eleonore

February 08, 2024
Tweet

More Decks by Eleonore

Other Decks in Programming

Transcript

  1. Who is this talk for? Anyone - all devs work

    with data github.com/Eleonore9/pyconse17
  2. Python data viz libraries “10 Useful Python Data Visualization Libraries...”

    seaborn ggplot Pygal Plotly geoplotlib Folium Meanwhile in the land of Javascript
  3. What should we make of it? Different libraries for different

    uses Here: dataset → quick viz → tell someone a story
  4. Wait, what’s Vega-Lite ? Vega-Lite is a concise JSON syntax

    for supporting rapid generation of visualizations to support analysis.
  5. So Altair is Simple API built on top of Vega-Life

    Aimed to produce beautiful visualizations with a minimal amount of code Note: Still at an early stage (v1.2.0) → more docs and plots to come
  6. Let’s have a closer look at Altair Installation (Python 3)

    For Conda users: $ conda install altair --channel conda-forge $ pip install altair $ pip install --upgrade notebook $ jupyter nbextension install --sys-prefix --py vega Note: To render Vega-Lite in a notebook → extension IPython Vega needed
  7. Let’s have a closer look at Altair Installation (Python 2)

    $ mkvirtualenv --python=/usr/bin/python3.4 altair $ pip3 altair $ pip3 install --upgrade notebook $ jupyter nbextension install --sys-prefix --py vega To initialize this nbextension in the browser every time the notebook (or other app) loads: jupyter nbextension enable vega --py --sysprefix CLI output
  8. Let’s have a closer look at Altair Doc example –

    import Altair, get dataset $ jupyter nbextension enable vega --py --sys-prefix $ jupyter notebook
  9. Produce shareable plots Easy way → use Github pages Note:

    rename your html file to “index.html”
  10. Produce shareable plots Workflow 1. Explore dataset and create plot(s)

    2. Commit changes and push to Github 3. Share the URL https://<gh username>.github.io/<project name>
  11. How to add to our page Let’s look at Altair

    source code to_html function We can pass keyworded arguments
  12. How to add to our page Navbar Title Paragraph Paragraph

    References Visualisations Page wireframe New function arguments
  13. Let’s look back at the steps Dataset (transformation) Chart() Html

    file Plots (template) to_html() Shareable URL Git Github
  14. Looking at a dataset with Altair Ebola outbreaks before 2014

    Source: https://data.humdata.org/dataset/ebola-outbreaks-before-2014 Open platform for sharing humanitarian data, managed by the UN (OCHA)
  15. Tips Organise your repository: input datasets /notebooks / helpers /

    images or html outputs Watch out for the next release supporting Vega-Lite 2.0
  16. Pros & Cons Pros: simple, work cooperatively (Github), reuse your

    code and html templates Cons: no handling of geo data, depending on Github Improvements: interactivity (next release)
  17. Thank you @EleonoreMayola github.com/Eleonore9/pyconse17 Resources Docs: https://altair-viz.github.io/ https://altair-viz.github.io/documentation/displaying.html https://vega.github.io/vega-lite/usage/embed.html https://vega.github.io/vega-lite/examples/#interactive

    https://vega.github.io/vega-lite/usage/applications.html Data source: https://data.humdata.org/ Blogposts: https://blog.modeanalytics.com/python-data-visualization-libraries/ http://pbpython.com/altair-intro.html Emojis: emojipedia.org