22 January 2018
As data becomes increasingly important and increasingly abundant, sometimes it can be difficult to extract the correct information from it quickly and present it in the right way. Luckily there are more and more tools coming onto the market to make the life of an analyst even more fun and also to make data analysis more accessible to a wider audience. In this article we are going to zoom in on two tools: SAS Visual Analytics (SAS VA) and IBM Watson Analytics (IBM WA).
New tools mean that data analysis is no longer the preserve of analysts with technical skills in various programming languages such as SAS, SPSS, SQL, R and Python. In recent years, the range of data analysis software has grown. For example, there are now packages that can turn data into visualisation with a simple play-like click & of a mouse. SAS VA and IBM WA are major examples of this: not only are they able to visualise data quickly and easily, they can just as easily identify correlations and even unlock a certain amount of scope for predictive modelling.
SAS VA is SAS’s answer to the growing demand for analysis tools that don’t require prior knowledge of SAS code and that are capable of visualising data powerfully and effectively. SAS appears to have done a pretty good job in that endeavour, with a tool that apparently emphasises its versatility rather than focusing on data visualisation alone. This versatility means that the tool also provides scope for developing dashboards and running simple process-mining analyses (Sankey diagram). In addition, the extra Visual Statistics module allows you to add predictive modelling to the package. Although it is limited to regression and decision trees.
This tool actually offers two structures for visualising data: in the form of either Data Explorer or Design Reports. The former, as its name suggests, is primarily aimed at exploration without the user directly working towards a structural application of the visualisation. This is a hugely convenient way to explore and visualise the data. Even more convenient is the Automatic Diagram function, where SAS makes its own suggestion, based on selected data, of what type of diagram to use to display the data. The Sankey diagram is also a good example of how SAS is capable of powerfully visualising data with this tool.
If you want to develop a dashboard, you can choose Develop Reports. Once you have made your dashboard, you can view it and share it with others using the Report Viewer. One way to do this is by granting read-only access so that the people who will use the dashboard can log in and see for themselves how the dashboard has been designed. Other ways to export the dashboard are as a PDF, csv or Microsoft file (Sharepoint, PowerPoint, Excel and Word).
In addition to data exploration and visualisation, SAS VA (combined with SAS Visual Statistics) also allows you to perform more advanced analyses, such as making a decision tree or running a regression analysis. The advantage of this is that the tool can take away barriers with regards to predictive modelling. For analysts who have never done any predictive modelling before, this therefore represents a good introduction to the world of predictive statistics. The disadvantage is that it remains primarily explorative since the tool does not include data preparation and implementation of the model (outscoring). Therefore a major prerequisite in order to use these methods is that the dataset has already been prepared and is therefore ready for predictive modelling so that a user can extract the right information from the data.
Overall, SAS Visual Analytics is a great tool that offers many possibilities but which does lose out on depth in favour of versatility. It is also disappointing that the visualisation of some of the diagrams is rather low when you zoom out to see everything at the same time (e.g. decision trees become illegible and unintelligible when you fully zoom out because the names of variables and values disappear).
Watson is IBM’s supercomputer which is controlled by Artificial Intelligence which answers questions asked by the user. IBM Watson Analytics (WA) is a cloud application based on these same core principles. Rather than performing the analysis yourself, now you can leave it to the tool by entering your analysis question and leaving the analysis to IBM WA. Very innovative compared to other data visualisation tools! Another big advantage of IBM WA is that the basic version is free, which makes it very appealing to just try the tool.
Like SAS VA, IBM WA also provides possibilities for exploring, predicting and reporting data using a dashboard. In fact, these applications can’t just be used based on the innovative questioning principle. Just like other tools, you can also, for example, set out by selecting a visualisation and indicating what data you want to display in it.
One advantage of IBM WA is that the tool itself already analyses a whole load of things in the background by means of Discoveries. This is particularly convenient for analysts as it means you automatically navigate through a number of major insights from the data without having to make a visualisation each time for yourself. This also means you are constantly being triggered and might just stumble across insights that you wouldn’t have drawn yourself. For example, you immediately look at correlations with other variables so you easily generate more insights than with other tools with which you have to create each and every insight yourself. The intuitive algorithms on which IBM WA is built make the tool accessible to a wide audience, and that is one of the tool’s great strengths. Even if you don’t have much knowledge of analysis tools, you can already go a long way just by asking the right questions and leaving the analytical work to IBM WA. This can in turn be a disadvantage for “real” analysts as they might prefer to do their own research. Fortunately, that’s why IBM WA gives you scope to analyse the data for yourself.
The Predictive Modelling options are pretty limited in both IBM WA and SAS VA alike. Their main strength is gaining rapid insight into the key drivers explaining the target variables. In this area, SAS VA offers more metrics and model information than IBM WA does, whereas the latter is mostly limited to a summary of the most important input variables.
In turn, IBM WA offers slightly more scope for visualisation in dashboarding, whereby users can, for example, add illustrations and figures to present your visualisation a little more forcefully.
IBM WA, like SAS VA, is also intended for prepared datasets in order to extract the right information as quickly as possible. Before you get started with IBM WA, you therefore have to prepare the data in another tool first before you can obtain the right insights – just as you do with SAS VA. The disappointing thing is that neither of these tools represents a good, straightforward data preparation solution, meaning that the time you save to attain the right insights is limited. Moreover, a separate tool to do that will also entail further costs. Therefore neither of these tools represents a complete solution for the purposes of analysis, and that is unfortunate, as both of these tools are therefore mere add-ons to the existing tooling and therefore potentially less interesting.
21 February 2022
Plan your calendar free and make sure you have plenty of party clothes in your closet,... read more
29 April 2020
About the Award The field of data driven marketing is evolving fast. To honor these developments,... read more
7 June 2019
As a data analyst you want to provide clear cut insights for your end users, enabling... read more