Wow, another week of Makeover Monday has already flown by and there were 90 submissions as of Friday morning (CET). This week we looked at Domestic and International New Zealand Tourism Spend. I hope everyone enjoyed the challenge and had a chance to learn a bit about NZ as well as one of its main industries.
All of us are constantly learning and I would think that most of us also make an effort to improve our work and our processes continuously. That includes Andy and me as well and because of that, I included data definitions in the data files. The Excel spreadsheet contained a sheet with definitions, while the TDE had comments for each field to ensure people knew what they were working with. I hope you found this helpful for your analysis.
I noticed that more and more of the submissions now include a listing of the data source and the author’s details, which is great!
I also liked that a few vizzes this week included an info button, which, upon pressing, revealed the data assumptions and explanations taken from the original viz. This is helpful for those wanting more information and is an elegant way to provide the detail without cluttering up the viz.
Thirdly, there was an optional dataset with geospatial data. Not required as part of anyone’s submission, but I enjoyed seeing the creative solutions some people came up with, especially doing tiled maps for NZ, which I hadn’t seen before. If there ever was a time and place for experimenting, #MakeoverMonday certainly is a good opportunity…
There is still work to do
Like I said, we’re all still learning and there have been a few key issues we have seen this week.
This dataset included an index. The Regional Tourism Index has a baseline of 100 based on Tourism spending via electronic cards during an average month in the year 2008 (which was not very eventful and was also the start of this data collection by the Ministry of Business, Innovation and Employment). Each region in the dataset has achieved a certain index every month over the years.
Being an index, you cannot sum it up like $$$, but some people used a SUM() rather than AVG() aggregation for their visualisations. Nope, that makes no sense. AVG() is the way to go here.
That being said, however, people also ran into trouble when they averaged those averages, for example across regions. That can’t be done and will yield incorrect results. The only way this data can be used is at either an individual region level [Region] or at a total country level. For the total country you had to use the Total (All TLAs) member of the [Region] field.
For a thorough explanation on this, please check out Andy’s blog from this week.
Note: I thought very long and hard about whether to exclude the total from the dataset because I anticipated that some people wouldn’t notice it and come to incorrect conclusions. But I didn’t want to temper with the data by deleting useful values. And also, you guys need to do a bit of analysis yourself. Us pre-determining everything won’t help you learn or get better at this stuff.
Lesson here: Analyse first, build your dashboard later. Look at the dimension members you are using and see whether they make sense. There’s an easy way to do so, which I wrote about here.
Annual vs monthly analysis
The data was at a monthly level, from 2008 to 2016. But 2016 was not a complete year so if you compare regions at a yearly level it is best to exclude 2016 because it only contains a few months rather than all months like the other years.
In my own dashboard I included 2016 for the topline charts and excluded it at the monthly level when I saw it was an incomplete year. I forgot to go back to my top charts and remove 2016 there. Thankfully, Chris Love pointed out my mistake and I updated my viz accordingly.
I appreciate that people don’t necessarily have the time or enthusiasm to go back and fix mistakes, but I wish they would. Some of the submissions were great visually and content-wise and I would have loved to include them in this week’s favourites. But if people don’t correct their errors even when we point them out, then I can’t justify including them in the ‘best of the week’ summary, because one main objective of these favourites is to promote best practice. So please, if you make an error, it’s not a big deal, just go back and fix it and let us know.
I love how some submissions included visuals to enhance the story, but please remember to – as a minimum – state the source of the image (‘Google Images’ is not sufficient: give credit where credit it due). Here is a great post by Ryan Sleeper on the topic of using the work of others and I hope you find time to read it.
Thanks for reading through the key ‘gotchas’ from this week. Let’s finally get into my favourite vizzes from week 4.
What I like:
- A stunning picture that goes well with the heading, the chosen colour scheme and the viz itself. The 2008-2015 trend chart in the bottom right corner reflects the peaks and valleys of the mountains in the picture. It’s something I hoped someone would visualise in exactly this way, so I’m really glad Rodrigo pulled it off so well
- The filters are subtle but effective and help keep the charts neat and minimalist
- Overall, the charts are simple and easy to understand, even for non-sophisticated viewers
- Data and image sources as well as author details are included
What I like:
- Simplicity: I love how simple this viz is, while still being visually appealing
- Simple colour scheme with two distinct hues of green works well to show the different visitor types and reflects some of the associations people make with New Zealand being a ‘green’ country with lots of outdoor attractions
- Reference line with 2008 baseline helps identify the shifting patterns in Tourism spending over the years
- Data source is stated, as are the author’s details
What I like:
- The chart has me intrigued straight away, I want to find out what the colours and the individual circles mean, so I find this a very engaging viz
- The map and additional visitor type filter bring some interactivity
- The calendar at the top right is a neat way of showing seasonal trends.
- The info button reveals all the assumptions and explanations around the data, which is great for those who want to find out more
- Data sources and author details are included
What I like:
- Great use of filters to let the viewer explore the dataset through the viz in a lot more detail. As a viewer I can compare against the baseline or against the country average.
Alternatively, I can pick different types of regions, e.g. cities or rural areas and then select the individual regions within those categories.
- I love the tiled map at the bottom of the viz. It shows, for example the drop in Tourism spend in the Christchurch region following the 2011 earthquake which I think is very clever. The tiled map information is enhanced by the two charts on the right
- I like the line chart at the center of the page and think it’s neat that David shows all of the years as circles in columns for each months, which lets me easily understand how the year I have chosen in the filter compares to other years
- The data source is clearly stated