Cycle plots are exceptionally useful for analyzing seasonality in time series data. In this video, I walk you through two examples that show the limitations of traditional time series analysis and two examples of how cycle plots overcome these problems.  Download the workbook here.

For Makeover Monday week 16, Eva and I are hosting a #MakeoverMonday Live at Tableau HQ in London. The data and viz this week were provided by Jonathan Drummey and the Visualize No Malaria project.

What works well?

What could be improved?

My Goals

Ben Moss has been post a series of really interesting insights that he's been finding in Superstore Sales. In particular, I liked what he found out about Binders...

Anyone ever noticed a seasonal trend in the number of products sold in the binder category before? #SampleSuperstoreInsights @tableau pic.twitter.com/zmTJeli7Yv

— Ben Moss (@benjnmoss) October 15, 2017

This week's challenge builds upon Ben's work. First, you'll need this version of Superstore (which is different to Ben's but the same one I normally use for WW).

Here are your requirements:

1. Create a line chart of the rolling N week cumulative products sold since the first week each product sub-category was sold. In this case, a product sold is the count of the product names.
2. Allow the user to pick a sub-category to highlight. Make it blue and all others grey.
3. The sub-category highlighted should also be slightly thicker than all other sub-categories.
4. Allow the user to define the number of weeks over which to compute the smoothing. The user should only be able to enter values between 8 and 52.
5. The title and the y-axis should update dynamically based on the number of weeks entered for the smoothing.
6. Match the tooltip in the chart

UPDATE (May 2, 2013): I have removed 2013 from all charts.

---

I posted an article on my Facebook page the other day asking if this radar graphs about artic sea ice volume works.  The comments were mixed.

I think this radar graph is merely ok.  Radar graphs, in general, are hard to read because relationships and trends are not easily discerned.  Some of the issues I see with this graph include:

• It’s difficult to see the entire pattern over time.  I see the pattern is spiraling in, but does it change year by year.  Ask yourself this: How does 1999 compare with 2004?  It takes a lot of work. 
• It’s very hard to follow the ice volume labels around the chart.  I bet you found this as well in the example above.
• The author only included September.  Why?  The data is available by day.  Get the data here.
• Are there any seasonal patterns?  You can’t answer this question.

I created a couple of different views for your consideration in Tableau.  You can download the workbook here.

In my version, I’ve addressed all of the problems I mentioned above.

• In the upper left graph, you can easily see that the artic sea ice volume is trending down.  
• The upper right chart allows you to see two things:
  1. The seasonal patterns
  2. Comparisons by year:  The comparisons across years can be see because the years get darker as the data gets closer to 2013.  You can see the lines get darker as you look down the graph.
• The Daily Volume graph makes the cyclical patterns much, much clearer.

How would you visualize this data?  Do you agree that these line graphs work better than the radar graph?