Advanced analytics is one of the most widely used terms in data analytics within the past years – but what is it?

Advanced Analytics and Reporting

November 28, 2019

Epinion’s Advanced Analytics expert uncovers Epinion’s approach to Data mining and how it can help you on your business’ journey on going from descriptive analytics to predictive analytics.

Advanced analytics is one of the most widely used terms in data analytics within the past years and with good reason: Advanced analytics allows you to utilize and dive deeper into your data and predict the future. Advanced analytics is often used interchangeably with ”Big data”, and while the latter indicates that we’re dealing with data more varied in structure, with bigger volume and more speed, Advanced analytics is the use  of the latest tools and technologies within analytics such as datamining or neural network/AI to deal with this data.

In Epinion, we use Advanced Analytics to help our clients move from looking at the past to looking at the future, in other words; going from descriptive and diagnostic analyses to analyses involving predictive and prescriptive analysis in order to take one step into the future. “Today, more than 80% of business analytics are descriptive, but once you have enough data, you can use Advanced Analytics to start seeing patterns. You can then build a model based on these data – and once you have built a model, you can predict.”*

In Epinion, we work with four topics within Advanced Analytics to support companies and organizations: Digital data collection, Dashboard Reporting, Data mining and Predictive analytics. These focus areas are used as steppingstones to help clients go from descriptive to prescriptive. Common for the four areas is that Epinion focuses on people; e.g. customers, travelers, students, citizens and so on.

At the core, working with Advanced analytics is not so very different to working with any quantitative research in general

In our opinion, working with Advanced Analytics is based on the same core principles that is important when working with any quantitative research in general. You still need to:

  1. Clean and prepare the data (clean for outliers, missing values and prepare the data so it has the right structure)
  2. Make sure the quality of the data is good (without quality data, forget getting quality answers)
  3. Creativity (you can always use your data in more than one way, a good simple idea can easily outperform a very complicated model)

That is why people with a strong methodological background in quantitative research often will be likely to succeed with Advanced Analytics.

Using Data Mining as a steppingstone on the journey going from descriptive to predictive analytics

Data mining is a way to cope with ever-increasing datasets and be able to make sense of it all. As more and more data are available, the need for being able to understand these data also increases rapidly. Datamining is all about discovering new meaningful patterns and insights from large quantitative datasets. Datamining is usually feasible and used where our regular statistics tools cannot cope with the large amounts of data – or need the latest and greatest technology.

In Epinion, we typically use it as a step on the journey going from descriptive analytics to predictive analytics, where the data mining helps us understand our data and the possibilities within – and at its fullest extent to be the first step in developing a prediction model.

Therefore, we see Data Mining as a point-of-departure in going from access to insights to access to foresights. Based on our experience, some of the benefits you can get from Data mining are:

  • Understanding why something happened, so you can avoid it happening again (or if the opposite, how to trigger it!)
  • Combining data that you haven’t previously combined to generate new insights and to lead to predictions
  • Segmenting people into homogenous groups so you can differentiate and target actions and messages towards them
  • Taking the first step towards being able to predict the future for your customers, citizens, travelers, etc.

Working with Data Mining (or any Advanced Analytics project in general) often requires a different approach in terms of the process. Epinion recommends taking point-of-departure in a specific business problem and scoping your project after this while allowing for learning underway. Typically, this is done through iterative processes with involvement from clients as well.

Evaluating the online discourse on Norway’s engagement in the Columbian peace negotiations

A good example of how Data Mining can be utilized to make sense of a large pool of unstructured data is a project Epinion conducted for the Norwegian Agency for Development Cooperation (Norad). The case is interesting because it illustrates how you can use Data Mining to dive into five years of vast Twitter-data to discover insights such as attitudes and influencers.

Norad evaluates Norway’s foreign political engagements and wanted to evaluate Norway’s engagement in the Columbian peace negotiations. A part of this evaluation would be to evaluate the online public discourse concerning the Norway’s engagement.

They needed specific insights into the overall key topics in the public discourse as well as the attitude towards their engagement in general. Norad also needed to gain a thorough understanding of which key actors that was influencing the public discourse, and how the different actors related to each other on various key topics.

By analyzing five years of Twitter-data using Natural Language Processing, Epinion gave Norad essential insights. Specifically, Epinion did a sentiment-analysis that showed the attitude towards Norway was neutral in the beginning but grew increasingly positive. Using keyword extraction, we identified the key topics, and lastly using network analysis, we identified the top influencers.

As an indication of the novelty in the project type, Epinion had an article published in SAGE Journals that was done together with a counterpart from Norad, on using Big data in Evaluation projects using the beforementioned project as frame of reference.

*Source: Gartner’s four stages of the Data Analytics Maturity model from “descriptive” to “prescriptive”

Want to know more? 

If you are interested in getting help to starting your Advanced Analytics-journey, or you want to hear more about how we conduct Social Media-analysis, please reach out to Manager Torben Jakobsen: