For this week’s #MakeoverMonday we challenged you to visualize 5,000 years of solar eclipse data, provided by NASA. The resulting visualizations were a collection is informative, interactive, artistic and creatively designed stories about solar eclipses across the planet and I truly enjoyed every new tweet that came in with another viz.

In return for the community’s effort this week, I am calling out three really positive observations I made from the submissions and ten, yes ten, short and sharp lessons for everyone to take away.
At the end I’ve listed my favorites for this week.

 

ITERATING ON FEEDBACK

We give people brief feedback on Twitter where possible and then have a more comprehensive session during #MMVizReview. It’s so good to see people using that feedback and turning the suggestions into improvements of their viz. Some may take everything on board, others take a couple of points and work with that.

We really appreciate your efforts to improve and learn and it makes it infinitely more satisfying to give feedback when it is received and acted upon in some way. While it’s not fun when someone asks for feedback, we give some and nothing comes of it (not even an acknowledgement that they’ve seen our comments), the reactions to #MMVizReview have been outstanding and we appreciate that the efforts from our side are matched very equally with hard work by you to learn and push yourself further following our feedback.

Suraj Shah iterated following our feedback during #MMVizReview

Tristan Guillevin also took our #MMVizReview comments and changed his viz

TEACHING ME SOMETHING

When I posted the data for this week, I asked the community to teach me something because I don’t know much about about astronomy, solar eclipses and all the things that happen up there in the sky :-).

I enjoyed seeing a number of very informative vizzes that explained the concepts of Saros numbers, Gamma, the relationships between time, cycles and types of eeclipses and that also brought those to me in an easy to understand and very well illustrated viz.

Well done everyone who took the time to dive into the topic and share their knowledge with the wider community. I know it takes extra effort but such educational material is helpful and our friends at Tableau Public have been watching with a lot of interest to see the vizzes coming out of this week’s challenge.

Angie Chen created an excellent dashboard this week which contained a lot of educational information and taught me a lot of new knowledge. Thanks Angie for that!

FINDING PATTERNS AND INVESTIGATING THEM

This week’s data was great for exploring and made it possible to really use visual analysis as a way to find a story (rather than having a story in mind and working to back it up with data).

It was a chance to put different measures and dimensions into the view, swap axes, add fields to colour and size and see what happens. I found the story for my viz that way and I know many others did too.
It was interesting to see people identify patterns in the data and working with those. Not only did that result in informative vizzes but also stunning visuals.

 found an interesting pattern in the data and used it as the focus of his viz. Not only did he create a very striking visual, but his viz was also picked by Tableau Public as Viz of the Day for Monday. Well done, Marc! 

 created a viz that revealed Saros cycles in the data, which not only was the first time I really understood what they’re about, but also made for a really great visualization of the pattern throughout the series and over time.

In the intro I promised you ten lessons from this week’s challenge, so let’s get right into it.

 

LESSON 1: WHY BLACK?

A solar eclipse can only be witnessed during the day when the sun is visible and even during a total eclipse there is still some light so why default to a black dashboard?

I knew to expect black dashboards and they can work really well. It’s important to remember that colour perception changes with the background colour. A colour that looks bright on white may look dull on black.

When you use a black background, don’t just use the default Tableau colours as they are designed for white backgrounds. Play with different palettes and see what works best for the design you have in mind.

If your sheets and text boxes need additional background shading to be impactful enough, then maybe a black dashboard background isn’t the best choice because you end up with a cluttered viz with various background colours.

 

LESSON 2: REMOVE THE CLUTTER

Outlines, gridlines, borders – do you really need them? Anything you put on your viz or dashboard potentially competes for attention with the actual data. Sometimes gridlines can be very helpful in guiding the eye and providing context. Often they aren’t really necessary, though and the same goes for outlines and borders.

If you really want to use them to divide up your viz, then at least make them subtle, i.e. thin, light lines rather than boxy big frames.

When it comes to decluttering, be generous. Whatever isn’t needed should be removed. Let your audience focus on the data and the main message. Make it easy for them by not having distracting icons, boxes, etc. on your viz.

Michael Mixon always delivers vizzes that are clean and only show what is necessary, reducing the ‘noise’ that could distract from his message

LESSON 3: USE IMAGES AND ICONS ONLY WHEN NEEDED

There actually is no need to use images and icons simply for the sake of it. I know it’s tempting to work with images of the sun and the moon, with images of solar eclipses and various icons and shapes that depict what’s going on. This, when done well, can be impactful and engaging. When done not so well it can look clumsy and unpolished.

Yes, we all need to start somewhere and I don’t want to discourage anyone from trying something new. When you stand back and look at your viz critically, ask yourself whether any images and icons you have used really enhance your viz or if simplicity works better in communicating your message. Don’t feel pressured to use them just because others are doing it.

 

LESSON 4: PRESENT YOUR WORK IN THE BEST LIGHT

Did you cut off part of your viz when cropping the image? It’s a real shame to see vizzes that are incomplete, just because the screenshot wasn’t executed properly.

If you viz is too large for a single screenshot, then export or copy the image from Tableau Desktop, which generates a proper image file.
Otherwise, simply go into presentation mode and take a nice, tidy screenshot, capturing ALL the parts of your viz.
While you’re at it, make sure your text boxes are all rendered properly and text isn’t cut off.

 

LESSON 5: DOES YOUR VIZ NEED A MAP?

I get it, maps are great, they’re easy to create in Tableau and many other data visualization tools, everyone is drawn to them immediately and it’s cool to show data dotted on a world map, it feels so grown-up (at least that’s the buzz I get).

Do we always need a map when we have geographical data? I want to suggest that we can often go without maps and find other ways to communicate the information effectively. Don’t get me wrong, I love maps and I love using them and there are very cool map based vizzes out there.

In this week’s dataset it was possible, however, to show great patterns in the data based on latitude and longitude without actually puttting a map behind it.

Here is a great example of that by Daniel Caroli who visualized Latitude without putting a map behind it.

LESSON 6: PROVIDE CONTEXT

When someone looks at your viz for the very first time, will they be able to understand it’s key message?

If you don’t provide an explanation in text, your viz needs to be outstandingly good at communicating the insight simply through the patterns, colours, shapes and trends in the data displayed.
If you have a simple bar chart with values and axis labels but no explanation, your title needs to be straight to the point.

A lot of the time it is a better choice to simply include an explanatory subtitle or some annotations and BANs (Big Ass Numbers) to guide your audience and make it clear for them what the viz shows.

Don’t let them guess or struggle. That isn’t a very professional approach when communicating information.
Give us context and don’t make us ask for it first.

 

In his viz, Mike Cisneros provided context in a simple but very effective way. A clever and intriguing title, a question to clarify what the viz will be about, numbered, short and few instructions so the audience will quickly agree with his claim of it ‘being simple’ and then a viz that actually works to support his instructions with additional information.

 

LESSON 7: GIVE US YOUR VIEW OF THE DATA

Sure, you can copy existing definitions and call it a day. Your viz will be much more impactful, however, if you put your own spin on it. Write your own descriptions and explanations and, most importantly, answer the ‘so what?’.

What insights did you gain? What conclusions are you making? Share them with your audience and give them a chance to learn something new. If they don’t agree with your findings, it’s a good opportunity for a discussion.

Axel used the first section of his viz to make a point about the frequency of eclipses and with it gives us his opinion, which sets the tone for the viz that follows.

LESSON 8: ARE PACKED BUBBLES THE BEST REPRESENTATION OF YOUR DATA?

When we have a lot of data points, the temptation is to use a packed bubble chart to show EVERYTHING all at once. But is it the most sensible way to communicate information?

Packed bubbles often don’t really show much when all the data points are very similar in size. Sure, the chart may look cool with different colours, but is it useful? Unlikely. Try a histogram or a bar chart which often brings across the message much more clearly and easily.

LESSON 9: TELL ME HOW TO INTERACT

When we build interactive dashboards and data visualizations, it’s easy to forget that not everyone who looks at it necessarily knows how to use the interactivity to its full effect.

Yes, we as a Tableau community are used to clicking and hovering and exploring the data with our mouse as we go. When we embed our vizzes in blogs or share them on social media, we need to consider ‘lay people’ and those who haven’t used interactive data visualizations before and ensure they know exactly how to navigate through the data.

When you use parameters, filters and actions, it’s a good idea to include brief instructions so that your audience can easily find the right places to click, hover, select and filter.

Tom Pilgrem added clear interaction instructions to his viz so his audience knows how to investigate the data on their own and drive changes in the visual by changing the data range and hovering over eclipse types:

 

LESSON 10: ANSWER YOUR QUESTION

Andy and I often propose to people to use a question as their title because it is engaging and provides a focus for the viz. It’s great to see more and more people try this approach. What’s really important is that your viz, in one way or another, answers the question.

If you start with ‘How many solar eclipses occured in the 19th century’, make sure your viz provides the number in a prominent place or you could add some text to conclude your analysis and share the result.

If you ask your audience to find out something from your viz (which I think is a great way to encourage interaction and exploration), make sure that whatever you’re asking them to find is very easy to identify, either through colour, shape, size or highlighting in some form.

 

THIS WEEK’S FAVORITES

Author: 
Link: Tableau Public

What works well:

  • Colin included a great amount of information on his dashboard without making it complex or cluttered. I learned something and the layout was easy to follow
  • Great use of colour to differentiate eclipse types and guide me through the viz
  • Appropriate use of images, they serve a purpose and don’t distract from the data
  • Striking patterns in the data to support his statements about types of eclipses and where they occur
  • He uses latitude data without the need for a map. Excellent choice!

Author: Curtis Harris
Link: Tableau Public

What works well:

  • Great simple design which is much harder to achieve than it looks
  • Excellent use of colour (and a departure from black!)
  • Well executed interactivity through hover actions and parameters. This is engaging for the audience and really helps make the story of the viz a personal one
  • Great use of BANs
  • The reference lines in the timeline chart are really well done and the labels are great!
  • Well-formatted tooltips

Author: Steve Wood
Link: Tableau Public

What works well:

  • Steve produced an informative and educational dashboard that is visually striking and gets by completely without black background
  • Interesting Analysis
  • Great use of colour
  • Great tooltips

Author: Athan Mavrantonis
Link: Tableau Public

What works well:

  • Athan found a striking pattern and visualized it in an impactful way using colour and worksheet size to provide a frame, almost like artwork you’d hang on your wall
  • Excellent hover actions that bring out the pattern as you explore the data further, together with nicely formatted tooltips
  • He also explains what the pattern is all about and what Gamma means, another viz that taught me something.
  • This could easily be used as a printed poster in a school for physics class