PyBay 2016 Datavis with Bokeh

Slide 1

Slide 1 text

Bryan Van de Ven [email protected] @bigreddot PyBay 2016 Data Visualization with Bokeh https://speakerdeck.com/bryanv/pybay-2016-datavis-with-bokeh

Slide 2

Slide 2 text

No JavaScript Bokeh? https://bokeh.pydata.org/en/latest Interactive visualization, widgets, and tools For the browser, with or without a server Versatile and high-level graphics Streaming, dynamic, large data

Slide 3

Slide 3 text

No JavaScript No JavaScript Bokeh? https://bokeh.pydata.org/en/latest Interactive visualization, widgets, and tools For the browser, with or without a server Versatile and high-level graphics Streaming, dynamic, large data

Slide 4

Slide 4 text

Prologue

Slide 5

Slide 5 text

Prologue LOGO: Simple inputs, sophisticated outputs

Slide 6

Slide 6 text

Turtles All The Way Down

Slide 7

Slide 7 text

Turtles All The Way Down Note to self: future Bokeh app example idea

Slide 8

Slide 8 text

Prologue LOGO: Simple inputs, sophisticated outputs

Slide 9

Slide 9 text

Prologue LOGO: Simple inputs, sophisticated outputs Fractint: Making pictures from math

Slide 10

Slide 10 text

Fractals All The Way Down

Slide 11

Slide 11 text

Prologue LOGO: Simple inputs, sophisticated outputs Fractint: Making pictures from math

Slide 12

Slide 12 text

Prologue LOGO: Simple inputs, sophisticated outputs The Web?! mod_python and PIL Fractint: Making pictures from math

Slide 13

Slide 13 text

Prologue LOGO: Simple inputs, sophisticated outputs The Web?! mod_python and PIL VTK: My First OSS Fractint: Making pictures from math

Slide 14

Slide 14 text

Prologue LOGO: Simple inputs, sophisticated outputs The Web?! mod_python and PIL VTK: My First OSS Chaco: Interactive rich-client applications Fractint: Making pictures from math

Slide 15

Slide 15 text

No content

Slide 16

Slide 16 text

Prologue LOGO: Simple inputs, sophisticated outputs The Web?! mod_python and PIL VTK: My First OSS Chaco: Interactive rich-client applications Fractint: Making pictures from math

Slide 17

Slide 17 text

Prologue LOGO: Simple inputs, sophisticated outputs The Web?! mod_python and PIL VTK: My First OSS Chaco: Interactive rich-client applications Conda: Users pay attention Fractint: Making pictures from math

Slide 18

Slide 18 text

Prologue LOGO: Simple inputs, sophisticated outputs The Web?! mod_python and PIL VTK: My First OSS Chaco: Interactive rich-client applications Conda: Users pay attention Bokeh: All of the above Fractint: Making pictures from math

Slide 19

Slide 19 text

No content

Slide 20

Slide 20 text

Community Setting the Stage

Slide 21

Slide 21 text

Community Setting the Stage

Slide 22

Slide 22 text

Community Makes the Dream Work Damian Luke Bryan Fabio Mateusz Havoc Almar Peter Maggie Sarah Brendan Kevin

Slide 23

Slide 23 text

Community Makes the Dream Work Damian Luke Bryan Fabio Mateusz Havoc Almar Peter Maggie Sarah Brendan Kevin New faces too: !!!

Slide 24

Slide 24 text

Community Makes the Dream Work Damian Luke Bryan Fabio Mateusz Havoc Almar Peter Maggie Sarah Brendan Kevin Special Thanks: Claire Teng, Dennis Obrien, Julia Signell, Sarah Heim, Tim Snyder, Steven Howell, Andrea Ernlund, Eric Ma, Nikola Hardi, John Dimattio, James Guillochon, Chris Coutinho, and Brigita Sipocz. New faces too: !!!

Slide 25

Slide 25 text

Slide 26

Slide 26 text

Bokeh Today

Slide 27

Slide 27 text

JavaScript / TypeScript API Today

Slide 28

Slide 28 text

JavaScript / TypeScript API Today http://bokeh.pydata.org/en/latest/docs/user_guide/bokehjs.html

Slide 29

Slide 29 text

Improved Layout Options Today

Slide 30

Slide 30 text

Improved Layout Options Today

Slide 31

Slide 31 text

Simple (to write) Apps for Data Science Today

Slide 32

Slide 32 text

Concentrate on your work Server

Slide 33

Slide 33 text

Concentrate on your work Completely written in Python, no HTML or CSS or “webapp” coding Server

Slide 34

Slide 34 text

Concentrate on your work Completely written in Python, no HTML or CSS or “webapp” coding Server Simple python scripts, no special classes of frameworks

Slide 35

Slide 35 text

Concentrate on your work Completely written in Python, no HTML or CSS or “webapp” coding Server Simple python scripts, no special classes of frameworks Useful for exploratory analysis or sharing and publishing

Slide 36

Slide 36 text

Slide 37

Slide 37 text

Concentrate on your work Completely written in Python, no HTML or CSS or “webapp” coding Connect the full PyData stack to interactive web apps Server Simple python scripts, no special classes of frameworks Useful for exploratory analysis or sharing and publishing Automatically mirrors and synchronizes Python and browser state

Slide 38

Slide 38 text

Streaming Data Today Two new data source methods: .stream to append data incrementally to column ends .patch for random access updates anywhere

Slide 39

Slide 39 text

Custom Extensions Today http://bokeh.pydata.org/en/latest/docs/user_guide/extensions.html Python class interface JavaScript implementation Normal usage

Slide 40

Slide 40 text

Custom Extensions Today Add capability for…

Slide 41

Slide 41 text

Custom Extensions Today Add capability for… Wrapping different widgets

Slide 42

Slide 42 text

Custom Extensions Today Add capability for… Wrapping different widgets Adapting 3D JavaScript libraries

Slide 43

Slide 43 text

Custom Extensions Today Add capability for… Wrapping different widgets Adapting 3D JavaScript libraries Adding LaTeX labels

Slide 44

Slide 44 text

Custom Extensions Today Add capability for… Wrapping different widgets Adapting 3D JavaScript libraries Adding LaTeX labels Community doesn’t have to wait!

Slide 45

Slide 45 text

Putting It All Together Today

Slide 46

Slide 46 text

(For the Audio Detectives) Today

Slide 47

Slide 47 text

Not just for Python Today http://hafen.github.io/rbokeh Plays well with R ecosystem: HTMLwidget, RMarkdown…

Slide 48

Slide 48 text

Billions and Billions… Today

Slide 49

Slide 49 text

Billions and Billions… Today (of problems plotting large datasets)

Slide 50

Slide 50 text

Billions and Billions… Today Overplotting (of problems plotting large datasets)

Slide 51

Slide 51 text

Billions and Billions… Today Overplotting (of problems plotting large datasets) Saturation

Slide 52

Slide 52 text

Billions and Billions… Today Overplotting (of problems plotting large datasets) Saturation Undersaturation

Slide 53

Slide 53 text

Billions and Billions… Today Overplotting (of problems plotting large datasets) Saturation Undersaturation Binning problems

Slide 54

Slide 54 text

Datashader Today http://datashader.readthedocs.io/en/latest/ Flexible, conﬁgurable pipeline for automatic rendering

Slide 55

Slide 55 text

Datashader Today http://datashader.readthedocs.io/en/latest/ Flexible, conﬁgurable pipeline for automatic rendering Solves problems of overplotting, underplotting, saturation

Slide 56

Slide 56 text

Datashader Today http://datashader.readthedocs.io/en/latest/ Flexible, conﬁgurable pipeline for automatic rendering Solves problems of overplotting, underplotting, saturation Mitigates binning problems by affording interactivity across large datasets

Slide 57

Slide 57 text

Datashader Today http://datashader.readthedocs.io/en/latest/ Flexible, conﬁgurable pipeline for automatic rendering Statistical transforms are a ﬁrst-class aspect of visualization Solves problems of overplotting, underplotting, saturation Mitigates binning problems by affording interactivity across large datasets

Slide 58

Slide 58 text

Datashader Today http://datashader.readthedocs.io/en/latest/ An independent project under the Bokeh GitHub organization Flexible, conﬁgurable pipeline for automatic rendering Statistical transforms are a ﬁrst-class aspect of visualization Solves problems of overplotting, underplotting, saturation Mitigates binning problems by affording interactivity across large datasets

Slide 59

Slide 59 text

Speaking of Dask Today http://dask.pydata.org/en/latest/

Slide 60

Slide 60 text

Speaking of Dask Today http://dask.pydata.org/en/latest/ A pure Python library for parallel and distributed computing

Slide 61

Slide 61 text

Speaking of Dask Today http://dask.pydata.org/en/latest/ A pure Python library for parallel and distributed computing Comes with a Bokeh app for monitoring performance of distributed tasks

Slide 62

Slide 62 text

Speaking of Dask Today http://dask.pydata.org/en/latest/ A pure Python library for parallel and distributed computing Comes with a Bokeh app for monitoring performance of distributed tasks “Invaluable for understanding and optimizing distributed computing performance.”

Slide 63

Slide 63 text

Slide 64

Slide 64 text

To be continued…

Slide 65

Slide 65 text

• Arbitrary computed columns (e.g. color) To be continued…

Slide 66

Slide 66 text

• Arbitrary computed columns (e.g. color) • SVG output (Chrome Headless) To be continued…

Slide 67

Slide 67 text

• Arbitrary computed columns (e.g. color) • SVG output (Chrome Headless) • NumFOCUS fiscal sponsorship To be continued…

Slide 68

Slide 68 text

• Arbitrary computed columns (e.g. color) • SVG output (Chrome Headless) • NumFOCUS fiscal sponsorship • Colorbars! To be continued…

Slide 69

Slide 69 text

Tomorrow Better Late Than Never

Slide 70

Slide 70 text

To be continued…  (continued) • Arbitrary computed columns (e.g. color) • SVG output (Chrome Headless) • NumFOCUS fiscal sponsorship • Colorbars!

Slide 71

Slide 71 text

To be continued…  (continued) • Arbitrary computed columns (e.g. color) • SVG output (Chrome Headless) • NumFOCUS fiscal sponsorship • Colorbars! • Community repo for extensions

Slide 72

Slide 72 text

To be continued…  (continued) • Arbitrary computed columns (e.g. color) • SVG output (Chrome Headless) • NumFOCUS fiscal sponsorship • Colorbars! • Community repo for extensions • Bokeh “develop” mode

Slide 73

Slide 73 text

Tomorrow Bokeh Authoring https://www.youtube.com/watch?v=VOAXC7XUCto

Slide 74

Slide 74 text

Tomorrow Bokeh Authoring https://www.youtube.com/watch?v=VOAXC7XUCto

Slide 75

Slide 75 text

Slide 76

Slide 76 text

• Binary protocol for arrays To be continued…  (continued) • Arbitrary computed columns (e.g. color) • SVG output (Chrome Headless) • NumFOCUS fiscal sponsorship • Colorbars! • Community repo for extensions • Bokeh “develop” mode

Slide 77

Slide 77 text

• Binary protocol for arrays • JupyterLab integration To be continued…  (continued) • Arbitrary computed columns (e.g. color) • SVG output (Chrome Headless) • NumFOCUS fiscal sponsorship • Colorbars! • Community repo for extensions • Bokeh “develop” mode

Slide 78

Slide 78 text

• Binary protocol for arrays • JupyterLab integration • 1.0 To be continued…  (continued) • Arbitrary computed columns (e.g. color) • SVG output (Chrome Headless) • NumFOCUS fiscal sponsorship • Colorbars! • Community repo for extensions • Bokeh “develop” mode

Slide 79

Slide 79 text

Resources • GitHub: https://github.com/Bokeh/bokeh • Documentation: http://bokeh.pydata.org/en/latest • Example Apps: https://demo.bokehplots.com • Tutorials: http://nbviewer.jupyter.org/github/bokeh/bokeh-notebooks/blob/master/index.ipynb • Mailing List: https://groups.google.com/a/continuum.io/forum/#!forum/bokeh • Gitter Chat: https://gitter.im/bokeh/bokeh

Slide 80

Slide 80 text

Call to Action Install Bokeh and provide feedback • BSD license for everything (JS, Python, server) • https://github.com/Bokeh/bokeh

Slide 81

Slide 81 text

Call to Action Install Bokeh and provide feedback • BSD license for everything (JS, Python, server) • https://github.com/Bokeh/bokeh Engage us to work on custom apps and dashboards • Code integration, analytical and visualization expertise • Feedback helps the open source efforts

Slide 82

Slide 82 text

No content