The Data Analytics Playing Field
It can be taken as a given these days that any modern BI and Analytics tool that finds itself riding a Forrester Wave or boxed into a Gartner Quadrant is going to have some serious visualisation capabilities. The playing field is leveling out and key differentiators between tool-sets and vendors are becoming harder to pinpoint.
Which is exactly why the Insights Foundry team at Synergy chorused hold my beer (in four-part harmony no less) and went digging.
Simple and Intuitive
Users typically want a simple, intuitive interface and the ability to produce lovely things like this:
This seems a perfectly reasonable request. More advanced users, particularly those sourcing their own data, would add another perfectly reasonable request for there to be a long list of hip and sexy connectors.
That’s all well and good, but saying you can connect to something isn’t the same as actually connecting to something.
Putting the tools to the Test
Always up for a challenge, our intrepid insights team put four major players – Cognos Analytics, Power BI, Qlik Sense and Tableau – to the test.
The brief was simple. Use the tool of your choice to connect to a generic set of contact centre data on Google Big Query and build a couple of quick visualisations.
The execution was…varied, but after making you sit through close to 150 words without a picture let me proudly unveil the Synergy Insights Foundry Mystical Quarter-Square*
*no magic was performed, no quadrants were harmed
The results are in…
The highly scientific scoring system (vetted and approved by the Chief Officer of Digital & IT at SARS) reveals Qlik as a clear winner.The scoring system suggests that Power BI needs to be given protection from itself, or at the very least from the Microsoft App Store – more on that in a moment.
Connecting with ease to Google Big Query for Qlik and Tableau was extremely straightforward. Both require the user to authenticate a Google account via a pop-up web browser – Tableau just edges it as Qlik needs a small extra step to copy and paste an authorisation key but there was little to choose between them.
Cognos Analytics lost points for two key reasons:
It requires a service account to be set up in Google – a fairly laborious and often counter-intuitive process.
It also needs a somewhat mind-boggling connection string that looks something like:
In a point and click world, neither of these could be classed as “easy” for a typical user.
Power BI started promisingly but fell apart entirely at the point of 2-step authentication to Google. It was very good at the sending a verification message to a phone part of the process but not so good at the listening to the response and moving on part and so the authentication window steadfastly remained.
I can confirm that neither restarting Power BI, rebooting your machine or reinstalling Power BI gets rid of the problem. Prayer, sacrifice and interpretive dance are equally fruitless. In fact, it seems the only way to resolve the problem is to uninstall Power BI and then never ever click on the standard download button from the Microsoft Power BI website ever again.
Trust me, if you see this:
…you are doomed.
But, if you click advanced download options and manually download the relevant file, everything works like a charm. It also fixes random and seemingly impossible issues with connecting to Oracle.
Shame on you Microsoft. How can the same version of a product differ depending on download path?
I can’t attest to the quality of the connection because by the time I connected I had lost the will to live. If I’m brutally honest, I only awarded it points so that it would appear on the magic quadrant mystical quarter-square rather than off the scale.
And the winner is…
Cognos may have had a few hoops to jump through to connect, but once connected it worked perfectly. Qlik and Tableau were equally quick to bring up a data preview and so after edging the ease of connection category Tableau was the clear winner.
Except it wasn’t.
You see, Tableau doesn’t like time. If it sees a column in Google Big Query defined as TIME it pretends that it’s not there. No error. No column metadata but populated with NULLS. Just…nothing.
So, if you don’t know your expected data set you’d be none the wiser – which is a major flaw.
You can work around it by writing custom SQL and casting the TIME fields as strings but that’s a hack at best. Points were lost, trophies were hastily redistributed and Qlik got the top right sector all to itself.
It’s all about context
The key takeaway from this exercise is to be thorough in matching your decision criteria to the problems a tool-set needs to solve. One man’s showstopper is another man’s minor irritation so it’s down to context and use cases.
The whole truth is greater than the sum of the slideware. Let Synergy help you get to the answers you need.
Connect with us!
Post written by Nik Eveleigh, Principal BI Consultant, Synergy.
Headline picture courtesy of pixabay