What a cool week for Makeover Monday. Aside from the live events that happened in Sydney, we also welcomed a number of new community members who submitted their first visualizations this week.
The dataset contained aggregated Sydney ferry trips from July 2016 to February 2017 at a month level. Not a big dataset but once again the community showed that even with less than 100 rows they are able to create great visualizations, find insights, practice new techniques and tell data stories in visually compelling ways.
I actually spoke to Transport for NSW on Thursday (May 4) about this week’s Makeover Monday challenge to see what they thought of it and am pleased to let you know that they loved seeing all these visualizations of their data being created by the dataviz community. They had a couple of people attending our live event and have shared links and tweets internally to show what happens when you let a bunch of dataviz enthusiasts loose on a dataset. They thought it was a really cool initiative and will use the vizzes from the community as inspiration to be shared in their organisation. So well done to everyone who participated!
This week participants not only delivered many many makeovers, but also engaged in constructive discussions around the data, the topic of public transport as well as visual best practices.
Others asked for feedback on their vizzes, received comments, iterated, made changes and improved their dashboards that week.
One of them was Neil Richards.
And then there were people who collaborated with others, did a viz together and learned from one another.
That kind of spirit, willingness to learn and supportive and encouraging behaviour is exactly what makes this community great and I’m proud that it is continuously growing and getting better every week.
Thanks everyone who got involved!
Let’s also look at some lessons that we can take away from week 18.
LESSON 1: GO EASY ON THE COLOURS
Sydney Ferries have a green colour palette on their ferry route map, using a different shade of green for each line. So I naturally expected to see a lot of green in this week’s vizzes.
It is perfectly fine to use colour and a ‘corporate’ or client palette if one exists.
Colour can be very powerful and should be used sparingly. It can be an extremely effective way to highlight parts of your data, indicate special data points that are particularly interesting, noteworthy or maybe even questionable.
If we use lots of colours on a viz it dilutes the impact any of those colours can have. It confuses our audience (‘what bit is important here???’) and can significantly lessen the impact our viz can have rather than enhancing it.
- Use colour sparingly. I often move away from the default blue that Tableau applies to charts and replace it with grey. More subtle, but still shows my bars and lines effectively. You can then work with shades of that colour to highlight particular dimensions or introduce a strong contrasting colour that calls our one specific aspect of your viz
- Think about the types of colours you choose. Do they trigger certain emotions? Are they associated with ‘good/bad’, certain brands or topics, etc.? Do they fit your topic, i.e. do they make sense for the audience? For example you wouldn’t use a creamy pastell pink when talking about youth unemployment or select yellows and greens when visualizing data on the US parties in the presidential election. Consider whether the colour palette you have chosen makes it easy to relate to the topic.
For this week’s data the colours that people would most likely associate with the topic of ferries would be shades of blue and green in reference to the ocean and the colour palette of Sydney Ferries.In her viz Pooja blends those hues through the image and background colours:
- Less is more. Take colour away and see whether your viz still tells a great story. If it does, keep it nice and simply. If it loses context, add colour where required
- Consider contrast. Black and white, grey and red, light blue and dark blue, there are some great contrasting colours out there. By using colours with a strong contrast you will not only make your data stand out but can help your audience comprehend your message faster and easier.
LESSON 2: WALK BEFORE YOU RUN
After weighing up the pro’s and con’s of providing images for people to use in their viz, I did decide to go ahead so that whoever wanted to play with design elements could use them in the process.
But just because we can doesn’t mean we should. Visualizing data in charts and creating a compelling visual story are not the same thing and I would caution people from going too far down the design end of the continuum before they’ve become proficient at communicating clearly with data.
What do I mean by that? Let me give you an example from my own history. I am not a graphic designer, and haven’t had any training in that area. In my early Makeover Monday submissions last year I wanted to do cool things like other people seemed to accomplish so effortlessly. But I didn’t know how to do it and the results were embarrasing…
So instead of focusing on design above everything else, I stick to improving my data stories and to getting better at communicating effectively through visualising data.
I don’t want to tell anyone not to try things out, because that is definitely what MakeoverMonday lets you do every week with new data. But be honest with yourself. If you don’t really like what you created, maybe have another go at it, simplify, and stay away from complexity if your data story isn’t clear.
- Just because there are a number of people who are exceptional when it comes to the visual design of their dashboards (layout, colours, images, fonts and font sizes, annotations and tooltips), doesn’t mean you have to nail all of the components on your first try. Keep it simple and focus on the data and showing your insights clearly before you venture into design territory. There is nothing wrong with a white background and you don’t need to include icons, images and floating elements just for the sake of it
- If you really desperately want to throw all your paintbrushes at the data (so to speak), consider creating a couple of dashboards. One simple version that just focuses on the data and the inisghts, and one that includes more design elements. At the end evaluate them critically and publish the one you like better
- Get feedback from others. Show your work to your colleagues, friends or family. Ask them what their first impression is, whether they understand your message quickly and clearly and what they think of the colour choices.
- Walk before you run. Get really good at creating a simple viz and then start introducing additional design elements. Also ask those whose style you like whether they can give you some tips on how to get started.
LESSON 3: PACKED BUBBLES
To be honest, I thought we had left packed bubbles behind, but they have reared their colourful, round heads again and seem to be making a resurgence in #MakeoverMonday submissions this week. Why? I’m not sure.
Andy and I have written about why we SHOULD NOT USE packed bubble charts in our recap posts for week 9, week 10, and week 12. Steve Wexler has written about it. Kelly Martin has written about it. And of course, the Big Book of Dashboards features a discussion on it.
So if you have used a packed bubble chart in your viz, I would ask you to review our earlier comments on the topic and to ask yourself whether a bar chart would show (much) more clearly what your data is about. Make it easy for your audience to understand the size, significance and impact of various dimension members of your dataset in comparison. Don’t make them look at bubbles.
Blow bubbles with your kids or while swimming under water, pour bubbles into everyone’s glass at the next party. But don’t put them on a dashboard if you want to communicate clearly and effectively. Please.
- Close the ‘Show Me’ panel on the right hand side and consider first what it is you want to communicate
- Instead of creating a packed bubble chart to show size comparisons, why not start with a bar chart? You can change the design of your bar chart to make differences more obvious. One approach is to use colours, but you could also change the ‘size’, i.e. width of the bar. If I am comparing distances, for example, I prefer to make my bars rather ‘skinny’ and have the length of the bar resemble how ‘far’ things are in comparison. If I’m looking at data which contains the size of something, I prefer my bars to be wider because it creates a perception of area and surface.
- Bubbles can also be quite a space consuming visualisation. If you want to give your audience an easy way to navigate via dashboard actions, consider using squares or rectangles instead and placing them in a single row or column to save space.
With all that said, let’s look at
THIS WEEK’S FAVOURITES
- Simple, clean design with white space that gives the data room to stand out
- Robert limited the card types to the four majory categories and used images of the actual cards to make it easy for the audience to relate to the topic, recognise the category they fall into and to engage with the content
- The colours of the card images carry over into the line chart
- The tool tips are effectively formatted and provide additional context for the viewer
- A simple and well formatted map gives the audience another interactive element and enhances the information provided by relating trip data to the different ferry routes
- The circles on the map don’t just show the destinations for each ferry route, but are also coloured and sized dynamically according to the filter selected at the top
- A clean summary dashboard effectively communicating the data
- Plenty of white space that leaves room for the data but also provides the right background against which to read the titles and headings
- The design resembles a corporate performance report and the horizontal lines are a great way to divide up the sections of the dashboard
- Nice clean and minimal font. The capitalization of titles compared to the normal sentence case being used for headers works really well
- Great idea to use the combined month filter and colour legend. This and the minimal use of colour for highlighting key data points helps the audience quickly understand what the colours are used for.
- The sparklines and their repetition across ferry lines help the audience understand easily what is happening with ferry trips over time
Author: Gabriela Plucinska
- The map looks really nice and simple and provides a visually compelling backdrop for the data
- The colours are nice and bright against the map and a distinctly different from one another which makes them easy to differentiate
- I like that the bar chart labels are also effectively a colour legend and the circles with each ferry route number look really neat. I also think it’s a great addition to have a grey circle for the ‘hub’ of Circular Quay where all the ferries originate
- The size legend makes effective use of the space in the bottom left corner of the viz
- The font is neat and minimalist and complements the overall design really well
- You can rely on Michael to always deliver top notch work and here is another example
- Great blue and green colour scheme where the colours don’t compete for attention but complement each other and ensure consistency across the viz, including the map
- The interactivity in the dashboard means the audience can explore the dataset easily and not just learn about the ferry trips taken but also see where the individual ferry routes are on the map
- Michael went to the effort of finding and including additional data for each ferry route, which while not expected, adds great detail and interactivity to his viz
- The data literally takes center stage
- Great colour scheme with hues that work well together but still provide enough contrast
- Using questions for the headings helps engage the audience to explore the viz. It also means the viewers know quickly what to focus on
- Tooltips are well formatted and provide additional context
- The stacked area chart looks like waves in the ocean which adds a nice design element without taking the focus away from the data
- Sub-headings are used effectively to specify what measures are being used in the viz and how to read the chart
- Very clean overall design and great attention to detail