Visualization techniques

From SESLINK Wiki
Jump to navigation Jump to search

( Back to the main page.)

This is an open and hopefully ever growing collection of links to favorite tools (libraries, packages) and techniques that help us produce informative figures from complex models.

Plain, useful figures

First you should select the main environment from which you do your analytics and want to produce figures. We have collected a few libraries for each of them.

R

  • The main R graph gallery
  • GGplot - a famous engine to produce figures in R. Provides the opportunity to customize every detail.
  • Plotly for R - makes graphs available in a cloud and enables interactive collaboration
  • Another way to make the figures interactive R shiny
  • Some links to spatial analyses in R: Rspatial or Spatial data analysis. A good overview is available in the R help pages for the raster package, a five page summary of the most often used raster functions.

Python

  • The main python graph gallery
  • Seaborn - a widely used library for pretty scientific figures, next to
  • Matplotlib - the classical standard for matlab-alike figures + the tutorial
  • Plotly for python - makes graphs available in a cloud and enables interactive collaboration + the tutorial to integrate this in a jupyter notebook
  • A way for a 'shiny' figure production within notebooks are provided by widgets, and a way showing how to do this is here

Matlab

Tips and tricks for polishing

  • Export figures from the original environment in vector graph format (svg, pdf ...) so that you can customize all elements by hand, which were not accessible via code. This helps also to align multiple figures of different types to produce one overview. An environment to load and adjust vector graphics is inkscape.
  • Particularly for maps, with different types of data and layers, it is suitable to export them individually (vector - svg, raster - png) and reassemble them in inkscape. In this way you can make sure that the files take only the least amount of space, you can still handle them and choose the specific layers for each plot export.

Infographics

This section is originally filled by notes from Nanda (thank you for sharing!) after visiting two FutureEarth webinars.

A. Take home points:

  • infographics are messages about our science. Communicated in a simple, attractive way that makes people want to read more or visit you site.
  • You and your brain = most important. Sit down and plan the thing through.
  • Prepare: do research, look at how others do things. Content: story, audience and medium needs to be known beforehand.
  • Design tips: love space, alignment and text - align everything physical to metaphorical. Use templates OR start with a grid to support alignment & consistency.
  • There are more tools, but these are ‘freemiam’ = free but not all parts of it

B. Tools to deal with data

  • Watson analytics: Ask questions to the data without having to know statistics; Steep learning curve
  • Tableau Public: Steep learning curve, Powerful statistics

C. Tools to construct an infographic

  • Timeline lab: easy, no maps, fill out some google sheets kind of thing and then it creates a time line
  • Infogram: simple, maps, cannot download in free account
  • Venngage: maps (mostly paid account),
  • piktochart: pay to download, but S version you can, no heat mapping, not scaling
  • visme: static version,but download is a little more relaxed
  • Canva: not really for the long form of infographics, useful for social media post - instagram, most functionality is available for the free content, pay for sharing etc

D. Tools for interactive graphs:

  • plot.ly: most technical
  • chartblocks: not rsponsive
  • googlesheets: mapping ability, not too big a pictures, print as pdf for big pic save (vector form..), not responsive
  • raw: gives access to a bit more weirder ways of presenting data
  • cartoDB: steep learning curve

E. Sources for inspiration:

General:

Specific (persons & infographics)

Misc:

Retrieved from "https://seslink.org/wiki/index.php?title=Visualization_techniques&oldid=624"