Note: Images may be inserted into your messages by uploading a file attachment (see "Manage Attachments"). Even though it doesn't appear when previewed, the image will appear at the end of your message once it is posted.
Register Latest Topics

  Author   Comment  

Posts: 4
Reply with quote  #1 
Is there an issue in using sparklines to compare data year on year but where you are only comparing part of the year i.e. Jan - March 2011, Jan - March 2012, Jan - March 2013 or Apr - Jun 2011, Apr - Jun 2012, Jan - March 2013 and so on?

I personally don't have an issue as I think compaing the seasonal differences justifies this but my boss disagrees (show Jan 2011 to March 2013 contiguously) and now I'm wondering if I was wrong in my assumption.

First time poster so apologies if this is outwith the spirit of the forum.

Posts: 805
Reply with quote  #2 

To make what you're asking clear, it would help if you posted an example.

Stephen Few

Posts: 4
Reply with quote  #3 
Apologies for the delayed response.

The example discussed would look something like the examples below. In the actual report the word Sparkline was replaced by Trendline.

January - March
Label A100908075 
Label B100110130160 
Label C10010510090 

January - June
Label A200180160150 
Label B170187221272 
Label C220231220198 

Posts: 4
Reply with quote  #4 
P.S. Looking forward to reading the new book.

Posts: 53
Reply with quote  #5 
Hi, CDevon. 

Lines are good for depicting trends, and sparklines are just small lines. So, if you want to look at how sales are trending over a specific period of each year, there's nothing inherently wrong with using sparklines for this. Whether this analysis makes sense for your business is something we won't be able to help you with, but if there's a seasonal cycle to the data it makes sense to me.

Stephen wrote a whole article on sparkline variations for different purposes (http://www.perceptualedge.com/articles/visual_business_intelligence/best_practices_for_scaling_sparklines.pdf) that you might want to check out. In general, though, sparklines are good for showing ups and downs. If you want to get a better sense of how the labels are doing relative to each other, though, you should consider plotting them all in one chart. This image shows both for your second table. (I've made the assumption these are sales numbers.)

Another view of the same data shows you the percentage of change in this period since 2010. Your labels all have numbers in a small range, but this is also a good way to compare trends among items with widely varying values:

My point is that there's no one right way to draw the graph, but that different views support different analyses.


Posts: 4
Reply with quote  #6 
Thanks for the feedback Jeff.

The issues was raised by a new boss that a trend couldn't be displayed in the manner of the tables above as it gave a false 'trend'. Personally I'd disagree as well as long as it is clear what data the sparkline is representing, but I was curious to flag it up on this forum.

Posts: 191
Reply with quote  #7 
It seems the bosses concern is that the chart might be interpreted as representing what happened between each period.

I think this is a valid concern, but I also think that the smallest effort in proper labeling and titles will clear any confusion up...

Previous Topic | Next Topic

Quick Navigation: