This week’s recap about the Tourism Density Index comes to you from Berlin, where I am playing tourist myself and enjoying a few days in the capital.

The dataset I provided was very simple with just 20 rows across 6 columns. 20 countries were ranked based on their level of over- or undertourism, defined by the ratio of locals to tourists.

I enjoyed seeing lots of vizzes this week and while I may have been more quiet than usual (we had our Exasol conference her in Berlin Mon-Wed, so I was happily enjoying the buzz around that), I did notice all the submissions coming in and particularly enjoyed the many simple bar charts. As usual, there is no need to over-complicate things.

During Viz Review, a couple of things stood out to me so I will turn those observations into my lessons for this week…




There are situations when you don’t want to synchronize your axes as they show a trend that wouldn’t be visible if the axes were synchronizes, e.g. when one measure is multiple times larger than the other, but they behave in a similar pattern over time.

In this dataset, however, we saw a number of visualizations which plotted the data as diverging bar charts but didn’t fix the axes, even though they showed the same metric.

For example, Sean Hurwitz created this viz with two bar charts where the bars have different values but the same length.

The bar for Croatia shows 14 as the label, while Tanzania has 43. They’re both the same length. This is confusing and his story for over- and under-tourism would actually be stronger if both axes were fixed, i.e. both go from 0 to 44, for example, to show the magnitude of over- and under-tourism for each of the countries.

While not too far off, Tushar More‘s axes are also of different length for the same metric and distort the comparison between under- and over-tourism, making the number of tourists for China (141 million) look the same as that for France (202 million).


Colors are a great way to draw attention to specific data points, to create a particular mood or connect different parts of a dashboard to name but a few purposes.

This week, I noticed that color wasn’t always used to the best effect and want to show a couple of examples from Viz Review and provide feedback on how these could be improved.


Positive versus negative

Using orange and blue as contrasting colors is perfectly fine. In this dataset, over-tourism is a negative thing, while under-tourism could be either, but is more on the ‘positive side’ as a situation that can be an opportunity for a country.

Vignesh Iyer created this diverging bar chart and applied orange to under-tourism and blue to over-tourism.

Aside from the different axes (see Lesson 1), I suggest swapping the colors, so that over-tourism (negative) is orange and under-tourism (rather positive) is blue, a more positive color.


Too many colors

The following donut chart by Ambarishan simply puts too many colors on the page. Aside from the fact that the data doesn’t add up to 10% and all the countries have the same sized wedge when neither their actual numbers nor their percentages are the same, the viz is VERY busy with all these colors and the colors change for the countries from left to right. No idea how to interpret this.


Background colors

Lastly, this viz by Chithresh Suresh is screaming at me with it’s bright yellow background. I wish the background was changed to a light grey so that I could focus on the data and the story instead. But the background color overpowers everything.


With colors there are a few points to remember and I recommend you start with these:

  • Less is more. Remove color where possible and focus on the important aspects of your story. Use color sparingly for these, e.g. have two colors in your viz (e.g. gray and red or gray and blue) with one being used for all the context and a strong color being used to highlight data points, insights, etc.
  • Consistency. Make sure your colors mean something and same colors mean the same thing, e.g. red for China in one chart means that the same red relates to China in all other charts
  • Choosing the right colors. Ask yourself whether the specific colors you have chosen will resonate with your audience. For example, typical ‘negative’ or ‘alarming’ colors in a Western context are red, orange, amber, potentially pink and purple. Positive colors are blues and greens. Keep this in mind when choosing your color palette.
  • Remove distractions. If you are using colors in your viz, even if it’s just two, be careful not to overpower them with a background color. I would personally only use a background color (other than grey) if my viz was monochrome. As soon as you introduce a strong color to highlight, e.g. red, blue, green, orange, etc., keep the background simple, i.e. white, grey, black or beige.


With all this said, let’s look at this week’s favorites…





Author: David Eldersveld
Link: PowerBI

Author: Luisa Bez
Link: Tableau Public