Inspired by the whitepaper Advanced Analytics with Tableau, I wrote this blogpost. Tableau strives to be accessible to a large part of the organization, bringing the amount of people who ‘can’ closer to the amount of people who ‘want’. In many Business Intelligence tools, technical proficiency is often the barrier for achieving truly advanced analytics and insights. Tableau itself uses the following picture:
It shows that conventional means gives a ratio of 8% between people, who want information, and people who can provide this information need. Tableau attempts to increase this ratio to 80%, making an organization much more self-reliant and less dependent on IT staff with specialist skills.
The goal is to give non-technical users, who are data analysts, access to advanced statistical capabilities and analysis to draw smart and process-improving conclusions. This makes Tableau a very important player in the field of “self-service-bi”.
The whitepaper, the inspiration for this blog, describes 5 advanced techniques that are available to a non-technical data analyst. These 5 techniques are:
I will take deep dive into these five topics using examples you will recognize.
Q: Is there a relationship between population, tourism and life expectancy of people in a country?
The dashboard above simulates a number of variables how the results will react to them. It’s possible to then take pictures of different scenarios (story-points) so that the analysis of certain combinations can be saved for later evaluation.
A key question in the world of Business Intelligence is often “what-if?” Doing analysis within the company that can help with this insight is of great importance. Especially if they can be tested against reality over time. Functionality in Tableau that can help with this can be found in:
Q: What would be the impact on sales performance when changing with the salary and commission of sales representatives?
Above, a dashboard with a number of variables simulates how the results will react to them. It is then possible to take pictures of different scenarios (story-points) so that the analysis of certain combinations can be saved for later evaluation.
In Tableau, there are four ways of applying advanced calculations to the source data. These are:
These calculations are a coherence of the dimensions against which data is accessed, aggregated and presented. This avoids apparent contradictions in these calculations and helps the user to present the right figures in the right way.
Q: How many new customers are brought in per region over time?
This LoD calculation avoids that existing customers are considered as new customers.
A holy grail in the world of Business Intelligence is proactive predictive insights. Among other things, this plays an important role in companies that rely on heavy machinery and other capital, which cost a lot of money when they are idle, broken and in need of repair. The ability to properly predict preventive maintenance is critical and competitive in this regard. Using time series, insight can be gained into seasonal fluctuations, trends and forecasts, among other things. Tableau offers the following tools for this purpose:
Q: How will the stock price of different companies behave?
Tableau has a lot of predictive models and statistical perspectives that are calculated simultaneously. It is up to the user which of these provides the best insight, but each outcome is easily derived and explained by choosing additional details. A user does not need to be a programmer to do this, but can simply choose a trend line and find various supporting numbers (such as probability, distribution and P-value) from it.
With the development of R (a statistical programming language, which builds on S, and is an open-source set of advanced statistical algorithms), a Tableau user gains the ability to apply these calculations in Tableau. It is going too far for this article to go deep into this specialized field, but econometricians and statisticians are familiar with it. Tableau features integration capabilities with programming languages such as Python, R and MATLAB to present the results of very complex models in a simple and insightful way.
Tableau, in my opinion, stands alone in the universe of Business Intelligence tools and analytics platforms. The way businesses can unlock their analytical functions is very sophisticated, very approachable and ingeniously thought out and developed. User intuition is paramount and my personal experience with Tableau confirms this. In my work we like to use the expression “a fool with a tool is still a fool”. A person responsible for business process analysis knows what he or she wants. Tableau offers the possibilities to be very self-reliant in meeting your own information needs.
Would you like to learn more about advanced analytics in Tableau? In this whitepaper (20 pages) we will take an even deeper dive into topics like:
Als Senior BI Consultant bij Cadran Analytics, is Rick expert op het gebied van Tableau.
