BACK

2018-03-29

- By Adam Nagus

Data and analytics has never been short on terms and acronyms to describe how data could or should be used within enterprise organisations. I have been part of Management Information (MI) and Information Management (IM) projects and run Business Intelligence (BI) and reporting teams for the last 15 years.

Over the last five years I attempted to evolve how Business Intelligence is implemented through experimentation, research and lots of support from some very passionate and clever people.

Business Intelligence has been a technological and data-focused process that has provided me with many skills and lessons, however the biggest frustration has always been the lack of value and excitement business users seem to gain from the hard work delivered by my teams.

Advanced Visualisation

Born out of a deep passion for finding new insight and playing with new toys, I founded a team to focus primarily on new tools. We called this first iteration of our group the "advanced data visualisation" or "advanced viz" for short. "Advanced Visualisation", put bluntly, was an attempt to shake off the shackles of the standard processes of delivering static reports, from platforms we had implemented using technologies that had been around for almost as long as some of my team members have been alive.

SAP Business Objects, IBM Cognos, Microsoft SSRS and Microstrategy are a few of the typical BI technologies most people will be familiar with. Our new group was looking at primarily QlikView from a company called Qlik (previously known as Qlik Tech). The major difference between QlikView and the BI stack vendors were:

  • The focus on visualisation over tabular reports
  • A flat architecture which removed the requirements for separate data integration tools (ETL), separate database or data warehouse, no universe or semantic layer.
  • In-memory analytics engine which outperforms most traditional database technologies

Having a flat simplified architecture was a huge game changer. I could now change the way projects were setup and delivered. We now had the opportunities to adapt to a new way of working which took advantage of QlikView's ability to take data and perform analytics on the information, without having to wait weeks for data extracts to be built, data models or changes to existing data warehouses to be completed, before we could start interrogating the data and doing some very fast rapid data discovery.

Getting back to the premise for this article, which is data viz vs visual analytics, just focusing on advanced data visualisation tools provides a quick way of creating charts and graphs on top of pivot tables. However, it doesn't provide a step change to the business users, as they are still getting descriptive analytics information and KPIs. This is what I mean by data visualisation.

Data Visualisation: converting data into charts, graphs and shapes to aid human users in detecting trends and patterns in their datasets.

What was missing from data visualisation was the intelligence, user experience and workflow we now demand and expect from mobile applications or advanced web apps. Having a scatter plot or line graph is useful to some users, however it doesn't provide the user with additional value or tools.

A great example I like to use is a ticket reservation application like Ticketmaster. If you downloaded and installed the Ticketmaster app on your mobile phone and it only provided you with a number of seats available for a particular performance, you wouldn't be very impressed or pleased with this application. As a user you would expect the following:

  • A visual of where there are available seats
  • Ability to search for availability for different performances
  • Filter on price and quality
  • E-commerce workflow built into the same application, which allows you to purchase the tickets

When we talk about visual analytics, we are looking for a similar experience for a business user using analytics. A user should have an intuitive user interface (UI) which provides them with information relevant to their role and includes a user experience which allows them to do root-cause analysis, issue resolution and define what actions to take, without having to open up different systems and screens to take steps based on their analysis. Visual Analytics is a methodology and process that focuses on using user-centric design principles, to create a guided application for a specific business problem, instead of creating a report based on a list of requirements.

Use-case driven assessments focus on creating a solution which integrates with an existing workflow to aid the user, instead of providing them with a report, based on requirements they have passed to a technical team, in the hope that the information they have requested will be useful.

Visual Analytics

In 2013, I rebranded our group from "Advanced Viz" to "Visual Analytics" in order to create conversation around what I hoped Visual Analytics could become and not just focus on data visualisation. A year later, Qlik and Tableau also adopted visual analytics as a term to describe their tools. I remember seeing the new branding at the Gartner Analytics event in London in 2014 and was pleased that the new technology companies were moving away from being associated with BI or calling themselves "next gen BI".

Working with Qlik and Tableau, I developed a practice of young talented people, who had the skills to convert a business problem into a guided application. This meant having to rip up the old BI job descriptions and create a new role which demanded:

  • Business Analysis experience
  • Design Skills
  • Industry knowledge
  • Advanced Analytics
  • User Experience
  • Agile delivery
Finding people with the above skills just wasn't possible, so an academy and curriculum was required to create visual analytics consultants.

What about all the deep technical data skills? Data modelling and data integration skills are still very important; however, using tools like Qlik Sense, Looker, Tableau, etc has meant that there is less reliance on requiring a vast enterprise data warehouse in order to develop something for the organisation which is able to add value in days or weeks. There is definitely still a requirement for BI skills; however, combining a large data project with visual analytics will provide your organisation with value and insight very early in the delivery cycle. The BI and Data warehouse skills will provide you with the foundation to scale the visual analytics apps across your entire organisation and customer base.

To Summarise

"Visual Analytics" focuses on creating a capability that either enables business users to create applications to be used every day in their role (self-service) or create a guided app, by a specialised team of visual analytics experts, which is provided to a group of business users. VA is a capability which creates applications with an intuitive UI and sensible UX that includes intelligence such as:

  • "What if" analysis
  • Forecasting
  • Issue resolution
  • Root-cause analysis
  • Next best action
VA is reliant on data visualisation but it goes beyond the visuals in order to create an end to end experience that focuses on value for the user.

To see examples of visual analytics applications, please check out some of our public demos here.

In a future article we will be discussing the different type of visual analytics platforms. This will focus on creating guided apps for the business versus providing a platform and architecture for self-service visual analytics.

1 comments

Posted by hdporn on 2018-06-15

nice work

Please be aware that we will be processing your input data in accordance with our privacy policy.