Tuesday 16 August 2016


According to Wikipedia - "mashup is a song or composition created by blending two or more pre-recorded songs, usually by overlaying the vocal track of one song seamlessly over the"

In the Business Intelligence software space a mashup is quite similar. It could be blending 2 different datasets together or taking the capabilities of one product and blending it with another and producing brand new, even better content. Qlik Sense took this idea seriously when they decided to build their Qlik Sense product primarily on web technology frameworks. Whist they give you 'desktop' software as an executable, you are viewing a web page. A smart web page at that but essentially it is just a web page.

The greatness comes from the fortuitous timing of making this move at a time when HTML5 capabilities mean that the ability to build cool looking sites with interactive visualisations to tell stories should be second nature. It is still a wonder to me that some websites are not built to be HTML5 and thus non-mobile friendly. You'll notice this if you ever go to a website on your phone and have to "zoom in" or even make parts of the page smaller because it doesn't auto change to your device.

What it means is that Qlik Sense can build amazing dashboards that are mobile friendly straight out of the box.

Qlik showcases a couple of great examples which you can use for inspiration or just admiration and here's a few of my favourites

European Migration http://webapps.qlik.com/telegraph/europeanmigration/index.html
Click on the play button to see migration patterns come to life:

Marketing 360 http://webapps.qlik.com/marketing360/index.html#/dashboard
I really like this one because it uses more modern HTML and CSS frameworks (mainly Bootstrap) and the delicate touches make a big difference. Filters pop out and drop-downs menus have a slight animation style that gives the app an impressive tactile feel. Click around and try it out:

Global Beer Consumption
Pretty basic but it's about beer for goodness sake so pay attention!

The full list is here: http://webapps.qlik.com/ 

See you next time folks!

Monday 15 August 2016

It's been awhile

Wow. So, it's been eons since I last wrote here (May 2015 to be exact) and it has definitely been an interesting past few months. The busy schedule I've had has been well worth the time and effort though. I am still with the equity research firm QMG and we have had our own ups and downs as you would hear about from any startup company of note. A couple of exciting things happening in that space in the next few weeks so I'll keep you guys posted at a later stage.

What has been on my mind lately, and something that justifies my spending some time getting back on board, has been the urge to contribute. Sharing my journey and ideas in the world of data visualisation and analytics was the reason for starting to write and blog in the first place. I'll be the first to admit that it was with some ulterior purposes as well and at the time I started this I was in the software sales industry. Sadly, it made me push out the ideas of what my products could do somewhat beyond what was actually possible. In any case I had lofty ambitions for that firm and I wish them all the best, but deep down it is still a joy to me to be able to pick and choose tools with the freedom that comes from being on the customer side instead of the other end.

Anyway, enough ramble, time for more grunt. 

I've finally wrapped my head around Qlik Sense and it's capabilities and am well on the journey forward. They'e done so much more to the product in their latest iteration (version 3.0) and it's meant that we could finally move away from Qlikview and adopt it in our data services to clients. The key to this has been the way that they've made it easier to create your own journeys with the way you can build a dashboard or visual or web page. Qlikview (the old product) has always been to me, the Swisse army knife of the Business Intelligence (BI) industry. Not always flashy, not always the easiest tool to use, but the most capable all rounder. Qlik Sense is turning into just that tool as well an potentially even more so.

With Qlikview you had the ability to build pixel-perfect style applications but the platform was stuck in the past. This was mainly due to the limitations of the underlying build (C++) and with other products on the market becoming more HTML5 friendly, something had to give. Qlik Sense has finally gotten to the stage where it's easy for those who know a bit of Javascript, HTML and CSS to be able to build a page full of interactive visuals based on Qlik Sense. 

"Success comes from standing on the shoulders of giants." This is an often heard quote that I pay homage to every day in the work I do. For my success its been all about learning by copying and hacking apart existing builds that have elements I want to use and by trial and error I come out on the other end with continually improving products.

Our old product looked like this:

The ability was there to place objects wherever you wanted and some quite innovative popups could be made using hide/show capabilities but ultimately the key issues with "old" Qlikview was its inability to incorporate new technology. For example, the inability of this product to automatically switch between being desktop friendly to mobile meant that if you wanted to run off two devices you needed 2 versions. Forget about building for both. What would be more annoying to a user that opens up the page on their mobile only to have to zoom in to portions of the page at a time OR to the desktop user who has your visualisation show up on only a portion of their screen size. Not good!

Qlik Sense on the other hand is built as a webpage. The 'app' if you can really call it that has the technology in place that means I was able to load it up on an Amazon server (AWS) and do the edits from any browser without having to install the product on multiple desktops. Sure this means that I can't work whilst I'm on a flight (unless I port everything onto my desktop) but that's easy enough to setup and it forces me to be able to relax and enjoy something else like a good book or a Bloody Mary. 

Because it's built as a webpage it means that moving from desktop to mobile is easy as the product scales to fit the size of your screen. Here's an example of the desktop version in widescreen layout:

and here's what happens if you change the browser size to be thinner (like on a mobile):

It automatically scales!

How cool is that. I write no code, I adjust no settings, it just does it. Now that is smart!

Anyway it also means that innovations like Tableau 10's ability to choose what your app looks like on different screen sizes (but not automatically) look somewhat dated. Granted, Tableau is an amazing tool especially if you're new to the game of BI, but the the king for me in the work I do, is still Qlik. 

One more thing before I go which is interesting as I dive deeper into the world of data science is the ability to connect R server with Qlik Sense and integrate forecasting methods that used to be performed by running analysis in one product, passing it to the next for further analysis and then back into another visual tool to produce results. Cumbersome at times, this inability to automate did mean a lot of analysts were limited in what they could do. Now I can hook up my visuals to my database and to an R server and run advanced forecasting (in this case ARIMA) to 'predict' the direction of data I care about.

Anyway, enough for tonight. Stay tuned for more in the next few days. 

PS. Anyone interested in Qlik should take a look, not just at the usual marketing spiel you get from the companies homepage, but also at the blogs of their users.

http://www.askqv.com/ - is a great source of the best of these blogs and something I visit to keep up to date quite often.

Also there is - http://branch.qlik.com/#!/ - this is the place where a lot of further extending of the product occurs. Instead of just being able to use what comes out of the box with Qlik, the community of developers has contributed to creating further capabilities for the Sense product that go well beyond what the product developers could ever do. It is brilliant as you get contributions from a variety of sources around the globe and its a melting pot of ideas.


Sunday 10 May 2015

Data Visualisation Catalogue

Data Viz Catalogue

Here's an interesting blog that categorises all the various types of data visualisations by their type and purpose. There's nothing worse than seeing visualisations that don't convey the right message about their data or confuse the user. I think this catalogue will be very useful for anyone looking to tell a story or get a point across. The more that data visualisations get used in business and social environments, it's important to understand what's going on underneath rather than just on the surface.

Speaking of proper visuals, this reminds me of the work by Jon Schwabisch over at Policy Viz. His blog focuses on data visualisation in the government space but he did launch a Remakes page some time ago. Basically people send in visuals they see in the news or media that have some flaw in them and Jon goes about remaking them to better get the messages across. Try it here.

Very interesting concepts and certainly things that I'll use in my every day work and you should too.

Saturday 11 April 2015

IBM Watson - the future is here

Hi folks, I've recently had some time away from data mining work (analysing statistical datasets and figuring out what's going on in terms of production and capacity in various economies overseas) and had a chance to investigate more cutting edge technology from around the web.

Not sure if people remember Watson, the $3m IBM machine that was built to 'think' like a human - cognitive reasoning and unrivalled processing capacity all trapped inside a box...

It's also the brains behind the Big Blue IBM machine that beat Kasparov at chess. If you don't know Kasparov, I'd suggest you google him.

In any case, this level of computing was something at the time, far out of the reach of many organisations large and small. However a tide is shifting and with further developments in cloud computing and more and more information available online, the time was right for Watson to get amongst it. 

What's on offer is a distributive type of service where IBM is looking for apps/projects to get involved in and as much as I may have been against the might of software giants fighting against the little guys, I'm not against it when they do so in such as  a way as to move away from traditional norms of expensive software, long build cycles and limited access.

IBM has already started recruiting companies to take part in their 

Built in the background of all this are certain API components that give rise to the level of smarts you see above.

Some such as Trade-Off Analytics give users quick ability to analyse static datasets 

Or how about this demo which makes conversational computing a reality like in this healthcare example - https://watsonhealthqa.mybluemix.net/

Traditionally, any kind of querying of data has involved either visually looking at data as a table or as a chart or digging into it manually.  Improvements have been made over time to codify and further automate the process however the process for users has always been about knowing how to use certain systems. These systems have been getting easier and easier to use but what better way to do this than what Watson is doing.

In the above example, I can search as if I am asking a question....  eg 'what are the causes of the common cold?', or 'how do I know if I have food poisoning?'

Watson reviews what you've typed based on speech software and then looks into it's vast database to find answers and not only that, it also gives a level of confidence to those answers as well.

It's been a while since I'd heard of Watson but I'm glad they're making a comeback and doing so in a manner that is more reachable than anything IBM-like I thought would ever be.

Stay tuned for more updates as I investigate this further, could useful in what I or many others involved in the analytics space do going forward.

Friday 6 March 2015

Big Data time

Been a while since my last post I know however I'm going to slowly get back into this. My recent venture back into the world of finance has given me much time to experience the highs and lows of what data can do to help improve the way we live, work and even play. There's a great video from TED recently about big data and the way it can help everyday citizens come up with solutions to what would otherwise have been very much ignored problems. Data and making it easy to access, use and understand leads to better story telling capabilities and whether it be for civic duties or for monetary purposes there is much that can be done to find these advantages. Anyway, check out this video from Ben Wellington - and maybe even check out your local open data website or statistic agency. There's information out there that can help you - you've just got to do a bit of digging.

Sunday 9 November 2014

New Job

So I've moved into a new role and will be leading up the Business Intelligence practice for an Equities Research firm here in Melbourne.

It's pretty exciting as I get to once again use my finance skills (more like brush the dust off of...) and get involved in the wonderful world of investments

For those of you who aren't sure, Equities Researchers have started forming up over the last couple of years since the GFC (Global Financial Crisis) as the rates of pay typically given in basis points has been in decline. It's only really those who offer great insights and ideas to fund managers who will be able to make any money in this current climate.

As for the company, QMG, we've got some pretty good insights into sector level analysis and we couple our information with data mashups which is what I've seen happening in other industries to get unique insights. Our insights will prove valuable to those not just in the investments sector but anyone really looking to make either strategic or market based moves.

Right now we are focusing on overseas markets as the ASX is not really the largest of all the global players (though we might add them on eventually however we do have some Technical Analysis provided on the local stocks).

In the meantime if anyone is interested in having a look at our products check us out here.

I'll be focused on creating an interactive and easy to use experience for anyone interested in the research we do.

It's going to be fun being on the buyer side for once!

Sunday 17 August 2014

Graduate unemployment and salary information

Saw a great article in the Australian Financial Review today all about the decline of graduate employment in Australia (original article here). Edmund Tadros has pulled together some great data on the salaries of students and percentage of those still seeking jobs after finishing university.

The presentation of data in news is something that is always intriguing to me and another opportunity to find ways to visualise the data and make it interactive for consumers.

Here's my quick 5 minute hack-up of the data pulled together using SuperDataHub.