Using Jupyter to Create a Community for Satellite Imagery Analysis - - PowerPoint PPT Presentation

using jupyter to create a community for satellite imagery
SMART_READER_LITE
LIVE PREVIEW

Using Jupyter to Create a Community for Satellite Imagery Analysis - - PowerPoint PPT Presentation

Using Jupyter to Create a Community for Satellite Imagery Analysis and Sharing DigitalGlobe Sean Gorman PhD. @seangorman Version 0.1 Our Jupyter Journey Version 1.0 What are we going to cover? 1. Our problem 3. Accessibility 5. Scaling 6.


slide-1
SLIDE 1

Using Jupyter to Create a Community for Satellite Imagery Analysis and Sharing

DigitalGlobe Sean Gorman PhD. @seangorman

slide-2
SLIDE 2

Version 0.1

Our Jupyter Journey

slide-3
SLIDE 3

Version 1.0

What are we going to cover?

  • 1. Our problem
  • 2. Why Jupyter
  • 3. Accessibility
  • 4. React UI/UX
  • 5. Scaling
  • 6. Community
slide-4
SLIDE 4
  • 1. The Problem
slide-5
SLIDE 5

DigitalGlobe collects 3,000,000 km2 a day and has a 17 petabyte archive of imagery

slide-6
SLIDE 6
slide-7
SLIDE 7
slide-8
SLIDE 8
  • 2. Jupyter
slide-9
SLIDE 9
slide-10
SLIDE 10
slide-11
SLIDE 11

Why Jupyter

  • Muhammad must go to the mountain

○ Don’t reinvent the wheel ○ Go to where the users are

  • Open Source
  • Go-to tooling for data scientists
  • Over 3 million users

Problem | Jupyter | Accessibility | UI/UX | Scaling | Community

slide-12
SLIDE 12

Extending Jupyter

  • jupyter-react - https://github.com/timbr-io/jupyter-react
  • timbr-machine - https://github.com/DigitalGlobe/timbr-machine
  • juno-magic - https://github.com/DigitalGlobe/juno-magic

Problem | Jupyter | Accessibility | UI/UX | Scaling | Community

slide-13
SLIDE 13
  • 3. Accessibility
slide-14
SLIDE 14
slide-15
SLIDE 15

Satellite Imagery is Big and Awkward

  • A high res scene of satellite imagery = 40 gb
  • Standard photo imagery is three band (red, green, blue = RGB)
  • DigitalGlobe has eight bands
  • Imagery needs to be atmospherically compensated
  • Imagery needs to be color balance (dynamic range adjustment)

Problem | Jupyter | Accessibility | UI/UX | Scaling | Community

slide-16
SLIDE 16
slide-17
SLIDE 17
slide-18
SLIDE 18
slide-19
SLIDE 19
slide-20
SLIDE 20
  • 4. UI/UX
slide-21
SLIDE 21

Customizing Juypyter for Domain Specific Use

  • Jupyter React as a framework
  • Customizing within Jupyter Notebooks
  • Customizing the organization of Jupyter Notebooks

Problem | Jupyter | Accessibility | UI/UX | Scaling | Community

slide-22
SLIDE 22
slide-23
SLIDE 23

Customizing Jupyter

slide-24
SLIDE 24

Organizing Jupyter

slide-25
SLIDE 25
slide-26
SLIDE 26
  • 5. Scaling
slide-27
SLIDE 27
slide-28
SLIDE 28
slide-29
SLIDE 29
slide-30
SLIDE 30
  • 5. Community
slide-31
SLIDE 31

Building Community with Notebooks

  • Notebooks publishing and role based access control
  • Notebook cloning
  • Community tier and open data

Problem | Jupyter | Accessibility | UI/UX | Scaling | Community

slide-32
SLIDE 32
slide-33
SLIDE 33
slide-34
SLIDE 34
slide-35
SLIDE 35

Ikonos, Sentinel2, Landsat, and Event Data

slide-36
SLIDE 36

DEMOS

slide-37
SLIDE 37

Thank you.

DigitalGlobe Sean Gorman PhD. @seangorman