two people looking at a computer screen

Three Advanced Analytics Trends Worth Knowing

Three Advanced Analytics Trends Worth Knowing

5 Minute Read

Using three practical examples, MNP’s Brian Foster discusses the near-term implications of artificial intelligence in data analytics.

​With today's never-ending tide of improvements in advanced analytics - and all the marketing that goes along with it - it can be difficult to decipher what's meaningful to your business and what is simply hype. Allow me to suggest three current trends in advanced analytics that should be on the radar of any CIO or analytics leader:

Artificial Intelligence (AI) is beginning to help automate routine Data Science tasks.

There is still a great deal of misunderstanding surrounding AI as a whole, but many of the most effective use cases are structured around helping knowledge workers do their jobs more effectively by automating routine tasks with greater efficiency and accuracy than humans can achieve. This is certainly true in Advanced Analytics, for example:

  • Founded in 2013, Tamr offers a patented software platform for enterprise-scale data unification that combines machine learning with human expertise. It is based on a concept they call "Agile Data Mastering", which seeks to solve the scalability challenges of traditional data warehousing.
  • DataRobot, founded in 2012, offers an automated machine-learning-as-a-platform solution; utilizing AI to select the optimal algorithm for a given predictive problem. Given the shortage of available data scientists in the job market, this seeks to solve an obvious problem for many businesses.
  • Since 2012, ThoughtSpot has been developing a search-based approach to self-service analytics through an AI-powered relational search mechanism that seeks to make advanced analytics and data integration more accessible to business users.

Expect more advances to come and the associated costs to slowly decline. This is an important trend that will continue.

Natural Language Processing (NLP) is helping grow Self-Service Analytics

For most of us, it's a lot more comfortable to ask questions in plain English (or French, etc.) than it is to use a graphical user interface or write code. Vendors, of course, recognize this and have been busy finding ways to provide exactly that:

  • IBM's Watson is famously known for its question answering capabilities that were featured on TV's Jeopardy!. Although the solution has grown considerably since then, NLP remains core to its functionality, allowing users to ask questions of the data in text form.
  • Tableau, probably the best known brand in self-service visualization, acquired start-up ClearGraph in late 2017 to enhance their NLP capabilities – both through smart data discovery and enhanced analysis.
  • Microsoft's Power BI now offers "Power BI Q&A," allowing users to ask questions of their datasets in text form and view automatically-created visualizations as a result.

These are just a few examples; many more will continue to emerge. My take is that NLP can be a great tool to assist in analytical discovery, but it will require a level of Data Governance found in more mature analytic practices. Time will tell if that assessment is fully accurate.

Deep Learning / Neural Networks 

Deep Learning now appears on Gartner's 2018 'Peak of Inflated Expectations" – and is expected to achieve "Plateau of Productivity" status in just two to five years. Still, this is an area where I'd encourage most business leaders to 'proceed-with-caution'.

Deep Learning (roughly, a prediction technique based on our understanding of how the human brain processes information) can certainly provide great value. It is now the standard technique for computer vision problems and tends to out-perform other methods with high dimensional data – but this is hardly the environment most small to mid-sized firms operate in.

For most prediction problems, we are primarily concerned about the accuracy of the prediction – not how new the utilized technique may be. Traditional machine learning algorithms – if it is fair to call them such – can still outperform Deep Learning in many cases where smaller data sets are in use. In short, I'd recommend that you focus first on the accuracy of your predictions, and consider all the tools available – even good old fashioned linear regression.

While there may be no slow down in sight to the stream of analytic advancements we're currently enjoying, the three points above should help you to focus on what is most important to your business situation. And, if you need help navigating the wide array of options available today, MNP is here to help.

Tomorrow's technology is shaping business today. To learn more about how MNP can help you can make Predictive Analytics work for you, contact Brian Foster at 204.336.6131 or [email protected] .


  • Performance

    December 05, 2023

    Highlights from the Federal Fall Economic Statement

    The Honourable Chrystia Freeland, Deputy Prime Minister and Minister of Finance delivered the federal government’s 2023 Fall Economic Statement (FES) on November 21, 2023.

  • December 04, 2023

    Considerations for reporting Asset Requirement Obligations by First Nations communities

    An asset retirement obligation (ARO) is the expected costs associated with the retirement of a tangible capital asset.

  • Confidence

    November 29, 2023

    How recent tax changes impact professionals and professional practice owners

    Explore the impact of recent tax changes on Canadian professionals and gain insights into planning ahead for professional practices.