What does Narratives for Qlik make of the 2016 Gartner Magic Quadrant for BI and Analytics Platforms?

By Barry Harmsen

Gartner Magic Quadrant for Business Intelligence and Analytics Platforms

The 2016 Gartner Magic Quadrant for Business Intelligence and Analytics Platforms was released last February 4. This year’s edition caused quite a stir on various blogs and social media platforms as most of the ‘old-school’ vendors were dropped from the Leaders quadrant, leaving only Tableau, Microsoft and Qlik as Leaders.

Many excellent blog posts appeared in the last week, providing further commentary and thoughts on the current state of the BI and Analytics market. Rather than adding another post with my own commentary, I will take a slightly different approach.

Narratives for Qlik

On January 20th 2016, Narratives for Qlik was released. In the press release it states that:

The Narratives for Qlik extension automatically generates dynamic narratives to accompany Qlik Sense visualizations, explaining what is most interesting and important from Qlik’s charts and graphs, thereby accelerating insights derived from data

I think this sounds very promising, so I decided to download the extension and try it out. What would Narratives for Qlik make of the Magic Quadrant?

First try

The first thing I did was enter all the scores from the 2010 – 2016 Magic Quadrants into an Excel file, translating the Ability to execute and Completeness of vision metrics into values ranging from 0 (minimum score) to 100 (maximum score). A score of 50 on both metrics means the vendor is placed in the exact center of the quadrant. (entering all this data was quite a chore, so in order to save others some time, feel free to download the Excel file).

Next, I loaded the data into Qlik Sense and created a scatter plot (borrowing the idea for color coding from Ross’ blog post at the Ometis blog)

Gartner Magic Quadrant for Business Intelligence and Analytics Platforms - Qlik Sense Scatter Plot

Then I dropped in the Narratives extension and was immediately met with this warning dialog:

Unpossible

Ok, so that doesn’t work. Let’s try something else.

Second try, line chart

As the Narratives extension doesn’t support scatter plots, I decided to create a line chart showing each vendor’s combined score through the years. The combined score was calculated by multiplying the scores for Ability to execute and Completeness of vision and map them to a 0 – 100 range. 100 being the best score possible, and 0 being the worst.

The result was plotted in a line chart:

Gartner Magic Quadrant for Business Intelligence and Analytics Platforms - Vendor Yearly Trends

When dropping in the extension a second time, it asked which chart it should use. After clicking the line chart an automatically generated narrative appeared. There’s a few settings that you can change (Verbosity and Story Format; Paragraph or Bullet List) but I just used the default settings:

Automatic narrative for the full chart

Puzzled look…

Oddly enough, while the narrative looked interesting overall, I was a bit puzzled by what was shown. Specifically, by which data had been included in the text. The narrative mentions only nine vendors, while there are 36 vendors in the data set. If there was a reason that only nine vendors were of interest, a reason for this choice, or a list of the nine vendors, was not provided.

Besides a general narrative, details for some vendors were also provided. Curiously, it only gave details for Microsoft, IBM and SAS, again not providing a reason for why it specifically picked these vendors. When right-clicking the narrative and choosing Export data the full data set (including all years and vendors) is shown, so the extension does seem to access all of the data in the target chart.

I feel that specifying why certain data was included (and other data excluded) could improve the usefulness, and user-friendliness, of this extension.

(The above being said, as I am typing this, it is 2:15 AM and it’s been a long day, so I am not in the most analytical mindset and might just be missing something 😉 )

Third time’s a charm

For my third attempt, I decided to limit the vendors just to this years Leaders; Tableau, Microsoft and Qlik. You can find the full narrative below:

Vendor Trends
Based on this selection, the chart measures Total score by Year and Vendor.

The three vendors increased from 2010 to 2016 with Tableau rising the most (178.5%) and Microsoft rising the least (6.07%) over that span. The three vendors all had the largest single decline, on a percentage basis, between 2014 and 2015. Of the three series, Qlik had the longest span of consecutive increases, increasing over three periods from 2010 to 2013. Tableau, with the higher relative standard deviation, experienced the most volatility while Qlik was the most stable.

For Microsoft:
Total score rose 6.07% (42 to 44.55) from 2010 to 2016. The largest single increase on a percentage basis, which was 47.47% from 30.21 to 44.55, occurred between 2015 and 2016. However, the largest single increase on an absolute basis, which was a rise of 15.61 from 43.52 to 59.13, occurred between 2012 and 2013. Total score saw a significant positive peak between 2012 and 2015, starting at 43.52, rising to 59.13 at 2013 and ending at 30.21. Total score saw a significant trough between 2013 and 2016, starting at 59.13, falling to 30.21 at 2015 and ending at 44.55. Values for Total score ranged from 30.21 (2015) to 59.13 (2013), averaging 44.43. The series showed the longest span of consistent decline over two periods from 2013 to 2015, falling 48.91%.

For Qlik:
Total score rose 29.27% (29.89 to 38.64) from 2010 to 2016. The largest single increase, which was a rise of 23.68% from 35.64 to 44.08 (+8.44), occurred between 2012 and 2013. Total score saw a significant positive peak between 2010 and 2015, starting at 29.89, rising to 44.08 at 2013 and ending at 32.94. Values for Total score ranged from 29.89 (2010) to 44.08 (2013), averaging 36.81. The series showed the longest span of consistent growth over three periods from 2010 to 2013, rising 47.47%. The series showed the longest span of consistent decline over two periods from 2013 to 2015, falling 25.27%.

For Tableau:
Total score rose 178.5% (13.77 to 38.35) from 2010 to 2016. The largest single increase on a percentage basis, which was 54.44% from 19.38 to 29.93, occurred between 2011 and 2012. However, the largest single increase on an absolute basis, which was a rise of 10.63 from 29.93 to 40.56, occurred between 2012 and 2013. Total score saw a significant positive peak between 2010 and 2016, starting at 13.77, rising to 44.02 at 2014 and ending at 38.35. Values for Total score ranged from 13.77 (2010) to 44.02 (2014), averaging 32.31. The series showed the longest span of consistent growth over four periods from 2010 to 2014, rising 219.68%. The series showed the longest span of consistent decline over two periods from 2014 to 2016, falling 12.88%.

 

So, what do you think?

Conclusion

Despite being somewhat puzzled by the Narratives for Qlik extension narrating only part of the data for larger selections (which might be user error on my part, I didn’t exactly RTFM before using it) I am quite impressed with the narratives that it generates. These narratives really do point out interesting things that aren’t immediately obvious when you are glancing a chart. I can definitely see myself using this extension in future projects.

Furthermore, the extensibility of the Qlik platform will hopefully mean that we will be seeing many more additional capabilities like the Narratives for Qlik extension in the coming time, and will hopefully ensure that Qlik again gets placed in the Leaders Quadrant next year!

Download Narratives for Qlik

Download the Magic Quadrant data in Excel

(will update later with downloadable application and share on Qlik Sense Cloud)

 


 

Just a reminder, don’t forget that the $300 early-bird discount for the Masters Summit ends on February 19th. Don’t miss it 😉

Masters Summit Italy - April 5 to 7 2016

About The Author

Barry Harmsen

Hi there, I'm Barry and I'm a Business Intelligence Consultant at Bitmetric and based in the Netherlands. Originally from a background of 'traditional' Data Warehousing, Business Intelligence and Performance Management, for the past few years I have been specializing in QlikView and a more user-centric form of BI. I have done numerous QlikView implementations in many different roles and industries. In 2012 I co-authored the book QlikView 11 for Developers. You can follow me on Twitter at @meneerharmsen.

6 Comments

  • 1
    February 12, 2016 - 14:54 | Permalink

    Hi Barry,

    We’d also been playing with Narratives this week, but hadn’t hooked it up with our MQ app. In fact I’d been using it across my solar power generation app (see future blog article!). It feels like it gives a starting point to picking out trends, peaks etc, but isn’t quite at the stage where you’d pop it in an app for non-technical / non-statistical end users to read through.

    By the way, if you or anyone else wants to see the Ometis version of the MQ app, please just email info -at- ometis.co.uk and we’ll share it with you.

    Cheers
    Alex

  • 2
    February 12, 2016 - 21:35 | Permalink

    Hi Barry (and Alex) –

    I’m a product manager at Narrative Science (the creator of Narratives for Qlik). Thanks so much for your post and for giving the extension a spin. I hadn’t thought about using the MQ data to create a sample app, but that’s a sweet idea.

    Anyways, I just wanted to personally respond to some of the issues you had using the extension.

    1) The reason that the narrative only talks about 9 of the 36 vendors isn’t because of their ‘interest’, but because the other lines for the other series aren’t complete. Since those data points are blank spaces, it becomes difficult to do comparisons across lines if the lines have a varying number of data points. The narrative should communicate that decision to the user and our next release will include that fix.

    2) You notice that the narrative only goes into more detail about three of the series. This is because the narrative only has so much real estate on the page and we didn’t want it to be too long (imagine if the chart had 100 lines!). Again, the narrative should definitely call out this decision. The three series are chosen based on having the highest average value across the length of the chart.

    We are releasing a new version of the extension in the next month that will include these fixes as well many more features. To Alex’s point, several features for this upcoming release and future releases are geared toward giving the user more control over the narrative they get.

    • 3
      February 14, 2016 - 12:07 | Permalink

      Hi Dan,

      Thanks for taking the time to respond.

      Your explanation of 1) makes perfect sense, thanks for pointing that out. I noticed there were incomplete series, but didn’t think there were that many (25 out of 36). After checking I see that there are only 9 vendors that appear in all 7 years. Quite an interesting observation in itself, so might be worthwhile to include in the narrative.

      Regarding 2) Couldn’t the verbosity slider impact that? I know the screen real estate is limited, but on the other hand it is a scrolling div, so could potentially show 100 individual items. I could see cases where I’d like to see all underlying items (such as in the MQ). Perhaps another slider ‘Max no of sub-narratives’ could be added? Just thinking out loud here of course, and I also understand that part of the appeal of this extension is that you can just drop it in without fiddling with a lot of settings.

      Overall, a very nice addition to the Qlik ecosystem, and one that I will be trying out across a lot more applications. Looking forward to the next release!

      Kind regards,
      Barry Harmsen

  • 4
    February 14, 2016 - 16:49 | Permalink

    Hi Dan,

    As Barry said, thanks for replying – always nice to receive such interesting and informative feedback. We actually showed Narratives for Qlik in a Qlik Sense demo to a prospect the other day and they loved it – so please keep up the good work! Looking forward to seeing the updates coming through.

    Cheers
    Alex

  • 5
    Gordon Shumway
    February 14, 2016 - 22:47 | Permalink

    You guys have spent more time analyzing this than the Gartner people spent creating it.

    • 6
      February 14, 2016 - 22:54 | Permalink

      Looks like someone didn’t find their favorite tool in the Leaders quadrant 😉

      Say Hi to Willie Tanner for me 😉

  • Leave a Reply

    Your email address will not be published. Required fields are marked *

    Read previous post:
    The Joy of QlikView Grid Layouts

    While most QlikView Designers are familiar with the works and principles of data visualization gurus such as Stephen Few and...

    Close