11

I'm interested to know what sort of software - beyond basic standard graphics out of Excel - people use to generate data visualisations.

Do you craft things by hand and then build them in something like InDesign? I'm fascinated by the process that changes a table of data into a visual esthetic and would also welcome reading recommendations.

Improve this question
2
  • 1
    This is quite a general question - an example of the sort of visual you're thinking of would probably get more useful answers. "Data visualisation" could mean anything from a single pie chart upwards.
    – e100
    Commented May 24, 2011 at 8:51
  • I did note that I was interested in "beyond the basic graphics of Excel" which I think is clearly beyond a single pie chart. However, I take your point and add that I'm interested in how more outstanding examples of infographics are generated. Thanks.
    – temptar
    Commented May 24, 2011 at 12:36

7 Answers 7

Reset to default
5

Adobe Illustrator actually has some very under-utilized capabilities to enhance chart representations of data. There's a good tutorial by Mordy Golding here, and his Lynda.com tutorials also go into this in excellent detail. For the kind of work I do, I'll use Illustrator in this way, or build things by hand.

For inspiration, and to give you an idea of how far data visualization can be taken with large datasets, check out this video of Hans Rosling's work. Search "data visualization" on YouTube for a bunch of other great examples.

Improve this answer
5
  • Here's an amazing typographic approach to information display, courtesy of Kelly Vaughan (Twitter: @documentgeek). tktype.com/chartwell.php
    – Alan Gilbertson
    Commented May 25, 2011 at 1:02
  • Thanks for that. I am coming to this from the side of someone who does a lot of photography related work and Illustrator is something I don't use all that much of. I will look into it further.
    – temptar
    Commented May 25, 2011 at 15:27
  • Illustrator can be a bit of a conceptual challenge if you're tuned in to Photoshop and aren't familiar with a vector-type program, so frustrating experiences with it may have kept you from using it. I'd recommend taking the time to study the basics, because it does offer many useful tools (even if you only use it as a sort of max-plug-in for Photoshop). Lynda.com, tv.adobe.com and anything with Mordy Golding's name attached to it are all good starting points.
    – Alan Gilbertson
    Commented May 25, 2011 at 21:28
  • @Alan Gilbertson: while seeing the first page of the "put the 'art' in chart" page ( adobe.com/designcenter/illustrator/articles/illcs2at_chart.html ) I saw the 3d pie chart which is (IMHO) not an example of good visualisation, as the 3D effect only distorts the proportions of the pie segments, as described by many autors writing about "how to lie with statistics" (D. Huff, E. Tufte, ...)
    – MostlyHarmless
    Commented May 28, 2011 at 22:15
  • The question was about software, not about specific imagery. Mordy's tutorial, similarly, is about what you can do with the software, not the science of infographics or data visualization.
    – Alan Gilbertson
    Commented May 29, 2011 at 0:21
4

If you work with a big amount of data I recommend you try gephi. It gives you nice control on what and how should be visualized.

Improve this answer
0
4

What problem are you trying to solve? The approach and therefore best tools depend on...

  1. Are you visualising data to (a) analyse it, explore it or open it up, or to (b) communicate a specific, known message about it?
  2. Who is your audience? In particular, are they (a) casual people who's interest you want to attract (e.g. readers of a magazine, people following a link on twitter, people who work in a related but culturally separate field), or (b) do they have a committed interest in this type of data (e.g. they look at this type of data for a living, or they're a keen fan of this sport...)

If your answer is 1a2a, you have the big challenge of making a data set enticing to casual passers by. Sketch up ideas (maybe with Adobe Illustrator, maybe pen and paper), swot up on or hire skills in interaction design and client-side web programming, then code something up in Raphael.js (my preferred option, works in everything from IE6 to iPads), D3 (great, popular option for proof of concept and prototyping but being purely SVG it doesn't work in IE8 or below, excluding about 40% of people), Processing (viewers must have Java, excludes about 35% of people) Processing.js (HTML 5 Canvas, which can be made to work with IE, and has some benefits, but limits interactivity somewhat), or, if futureproofing isn't an issue and you want to save development time and money, Flash.

If it's 1a2b, your needs are to quickly transform and experiment with data, then show it to people who will be interested. Attracting attention isn't a priority - they're already interested - the key is being able to respond quickly to new questions. Look at stats-y data suites with good visualisation features like JMP, Tableau, R, or if there's no budget and R is too hard, Many eyes, Gephi, or WEAVE.

If it's 1b2a, this is classic infographic design. Get your artist's hat on, and crack out a vector program (design initially in vectors for accuracy and flexibiity) like Adobe Illustrator (which also has basic charting tools), Corel Draw or Inkscape (free) - or, if you've got the time or budget, make an awesome 2D infographic video in Flash or After Effects or an extra awesome 3D video in Blender (free, hard to use), Cinema4D (expensive, moderately hard to use), or Maya (very expensive, hard to use, industry standard).

If it's 1b2b, relax, you're communicating a distinct message to people who are already interested. Speaking clearly and confidently would probably do the trick. Or just give them your Excel file. Or send them an email or PDF with a well designed bar chart or table in it. Don't be too flashy - your audience probably know the nuances of this type of data as well as you do, and they probably don't want a presentational layer between them and their data. Unless it's a complex, unfamiliar data set - in which case, see 1a2a or 1a2b.

Improve this answer
3

I see a lot about professional and commercial software here, so this one might be a bit off-side:

I use LaTeX and TikZ (which is a LaTeX package) for visualization. If I am able to draw and structure my data on a piece of paper, I can also do the same thing with TikZ. The approach is 100% text based, not at all intuitive to beginners, but very powerful.

Some examples of basic functionality can be found at http://www.texample.net/tikz/examples/.

Improve this answer
2
  • I wasn't familiar with TikZ - thanks for mentioning it.
    – temptar
    Commented May 25, 2011 at 15:26
  • There's a sister SE site dedicated to LaTeX, including TikZ at: tex.stackexchange.com
    – Saaru Lindestøkke
    Commented Oct 29, 2013 at 15:32
3

I mainly use R to visualize data. It has a myriad of packages that extend its use. For instance, see R Graphics Gallery.

Improve this answer
1

This is a very general question.

I use e. g. statistical software like SAS JMP ( http://www.jmp.com ) for data analysis and visualization. It has very powerful features for "explorative" data analysis.

However for a clean an concise layout of the graphs for scientific publications, which I'm writing with LaTeX, I'm using IPE ( http://ipe7.sourceforge.net/ ). I export the diagrams created with JMP (or other software) to a pdf, convert it with pdftoipe to an editable vector graphics and then can use IPE to modify it, use mathematical symbols and formulas in high typographic quality, etc.

There are many other interesting approaches for data visualization, e. g. Tulip http://tulip.labri.fr/TulipDrupal/ .

Improve this answer
0

For publication quality figures I always use matplotlib. It can produce vector as well as raster graphics and lets you change fonts and do all types of customization. It is based on python.

You can see the examples here

Improve this answer

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.