This is a side project next to our regular jobs as Assistant Professor at NYU (Enrico) and Independent Truth & Beauty Operator (Moritz). We started in 2012, and learned podcasting as went along — we just felt it would be great to have a regular conversation and share thoughts on the role data plays in our lives with people whose opinion we value!
3 years later we have a listenership in the thousands, over 60 episodes and many many more we want to record.
Some important topics we touched upon include:
On the show we had a quite a few of really amazing people, just to name a few:
As podcasting is fundamentally a broadcast medium (oldschool, we know ;) this is also a great way for us to get in touch with our mysterious listenership.
Ask us anything and let us know how we can improve the show or what/who you would like to hear. We are super curious for your thoughts and questions!
Other things you can ask us about:
Freelancing, working from home
Balancing family and work
Design vs. academia
We will be back at 1 PM ET to answer all of your questions.
We are here — answering your questions! Keep'em coming!
OK, we are outta here, for now — that was fun. Thanks!!
Where do you stand on the recent Stephen Few/David McCandless debate? Should data visualisation "inform effectively"?
link here: http://www.perceptualedge.com/blog/?p=2154
I think SF is making a category mistake; he seems to criticize David McCandless' works not from a pop culture perspective, but his BI view. To me this seems like complaining that the crime novel does not read enough like a police report. They work in different spheres and genres; if you respect that, one can learn a lot from each other.
You ever watch Chopped or Top Chef and someone uses truffle oil thinking its upscale but the judges all make exaggerated terrible faces and say "ew, yuck, that's never a good idea and you ruined it you unsophisticated yokel"? What's the truffle oil of data visualization?
Ha, nice one! Bonus points for the cuisine analogy!
No direct analogy, but I once wrote on the role of "bacon" in data visualization — in short, using cheap tricks like faux 3D and fully saturated colors to make something look flashy, but ultimately not really a healthy information diet :)
3D is really a thing people love to apply to make things look cool, but often dominates the experience and hides half of the data (think globe vs map…) so — use it only when super appropriate.
The other thing I am pretty tired of is "pointless circlification" — bar charts don't get more exciting just because you wrap them around a circle. It's a cheap trick to draw attention (our eyes are magically drawn to anything bursty or eye-like) but not really helpful or particularly interesting…
Virtual reality and data visualization, do you think they will be used together in the future?
First, one could think that immersion (i.e. shut off your analytical brain, and just sensually experience the moment) and analytical thinking (which requires more of a third person perspective?) maybe don't go that well together. .
On the other hand, it could be great to build some really complex landscapes of data to explore, if people spend enough time and maybe it's annotated / narrated well, it could be a whole new level of experience.
I find e.g. the prospect of 360 degree video super promising for documentaries.
You have interviewed tens of creators, developers and researchers, do you feel the field has enough diversity of ideas?, do you have examples of interesting diametral perspectives?
Thanks, great question!
No, in fact I think there is some massive group think effect in a sense that both the design as well the science side as well as the side business of datavis often seem quite uniform (inside each group) in terms of what they like or how they work.
I'd love to see more non-standard approaches!
Will think more about good examples for diametral approaches. Any suggestions from the community?
Who's really your favorite, Robert or Andy?
Tough one! Do I go for looks or brains… let me think about this a bit!
Can you remember a time where the use of statistics dramatically changed your opinion on something? A scenario where the stats disproved many of your preconceived notions about a topic?
When I heard that 89% of statistics are made up, that sort of made me think.
Is there a market/need for visualisation of smaller data? Or are you only working with hundred thousands and millions of datapoints?
Yes, absolutely! In the Data Cuisine project, we often create dishes representing only a couple of numbers; very small data, but when they express important data and you meditate on them while preparing and consuming the dishes, this can be a very satisfying and informative exercise.
In a way, they more you know about a dataset (and the stronger your personal position is towards it) the more you can boil it down to a manageable size.
But, all good data visualization is generous in a sense that it should always offer a little bit more than what you asked for. The Google/Siri way to answer your question for the unemployment rate is a single number; the datavis way is to show you the number, but also reference points and context (how has it developed, how are other countries doing, how does it compare to the GDP, …)
How do you think Big Data will impact us as a society, and what is your stance on privacy rights in this digital age?
In short, I have huge concerns about the digital consolidation into only a few big players, and the resulting concentration of data (=power) in the hands of very few, private/commercial actors. Data privacy and "informational self-determination" will be huge issues in the years moving forward. I don't know if you watch Black Mirror — in the last few months I was often reminded how close we actually are to the dystopian "science fiction" scenarios painted there.
What's the best way to start freelancing?
Do a few projects that show what you can, and most importantly, what you want to do. Be part of communities / meetups / … Share and help! Understand that 30%-50% of your time will not be billable — calculate accordingly. Learn to manage your time, and communicate effectively. Be good at emails and calendars. Make sure to keep the boundaries between work and free time intact.
That's all I can think of right now — I have been doing it for 10+ years and am still figuring it out :D
What's the best way to start freelancing?
One more thought - don't look for the perfect project, but, at all times have a combination of projects that together work well for you (e.g. one might be good for the wallet, another one good for the soul…) - don't chase unicorns!
I know d3, but I am terrible at making things/charts look nice (layout, colors, legends, interaction). All my charts look like XLS output. Where should I start to improve the aesthetics?
Typography is a huge part. Work through http://practicaltypography.com. Think about what people should see first, and what is peripheral "lookup on demand" information. Keep it simple and clean. Observe and apply.
What are the general rules or processes you follow when tackling a new set of data?
Always assume the data is biased, incorrect, lacking, lying in some form… Look for what's missing!
Create as many views as possible, with simple tools. Look for the recurring patterns. Get in a dialogue with the data. Don't narrow down too early. Go broad, with cheap means early on; narrow, with expensive means later on…
Is there any software you use specifically that helps you find the stories in the data?
I like to use Tableau for quick data analysis, especially tables, plots and maps. For network data, it's Gephi. I am also quick enough in python and d3 to quickly code something up. Other people use R or Excel for these early explorations. The most important thing is really to find an environment that allows to quickly produce a lot of different views on the data. If it's cumbersome, you will settle down too early and miss out on a better solution or more interesting story.
What, in your opinion, are best practices for showing vast amounts of data to people that aren't data scientists?
First, I think you shouldn't show "vast amounts of data" directly, but find good ways to structure, prioritize, juxtapose, sequence, nest, … interesting information. You can show rich visuals but they need visual hierarchy, and structure. For non-experts, be careful with language (too technical language easily scares people off), Look into narrative techniques such as stepwise reveal, overview.and-detail, and annotations… Generally, people can digest quite complex things in my experience if it's presented right.
What, of all the dataviz you've seen, would you consider to be the most beautiful?
What about the most effective or compelling?
My all time favorite are pretty much everything from Wattenberg/Viegas, Ben Fry, Jonathan Harris.
Do your clients approach you with their projects? Or do you approach clients? Or do you sometimes just start with a dataset and look who might be interested? How does this freelancing-thing work for you?
Good question. Usually clients approach me with a challenge. Often, they have some other project of mine they really liked and want to have something "along these lines" which can be a bad start. I prefer of they just outline the problem and give the good access to data and have good ideas on the strategy side of things, and leave tactics/design to me. I also tried to approach clients with concrete ideas, but actually, this rarely worked out! I also sometimes do self-initiated/self-commissioned works but then they usually end up as art/own projects, rarely turn into client work directly. In the end, the mix is important!
Can "data visualization" be an umbrella term that contains from visual statistics to data art, and in that sense, having different goals along the spectrum? Can we say that, in some areas, the primary goal of data vis is not necessarily to inform, as long as this is clear for the audience?
Sure, why not? Journalism encompasses all kinds of activities, and outputs, from click bait to think pieces, opinion. Film making ranges from documentary to fiction with all its sub genres. I think data visualization has lots of genres, too! And it can only advance the field to learn about these, and learn when to apply which method, or narrative stance… (see also Wattenberg and Viegas on data vis genres)
Love your podcast! A couple of random questions.
What advice would you give to 'intermediates' in data visualization? A lot of information is available to beginners, but there is not as much direction for people who are past their 'beginning' steps.
What kind of workflow do you use? I know that the tools are only as good as how you use them. What OS do you use, what do you use to clean, etc. Or is it to dependent on the data at hand?
Speaking of set ups, Moritz. Is this you? i3 config youtube video I swore it was, what does your set up look like? I just got into i3 myself
Yes, agreed. Connect with peers on the same level, practice, learn from other communities…
For me, it's usually Tableau, Gephi, Python, d3 in the early stages, custom code in coffeescript / ES2015 + all the rest of the web things or processing for the end product. Sublime Text.
Haha, no!! No idea who that guy is, but NOT ME :D
- t3_3rcwfe_comments.json 126 KB
This article and its reviews are distributed under the terms of the Creative Commons Attribution 4.0 International License, which permits unrestricted use, distribution, and redistribution in any medium, provided that the original author and source are credited.