“OUT-OF-THE-BOX” ADVANCED ANALYTICS IN DVCS

Oracle Data Visualization Desktop has the following “out-of-the-box” advanced analytics built in:

1. Reference Line
2. Trend Line
3. Forecast
4. Clusters
5. Outliers

The way you use the analytics and add them to your visualizations is the same for all of them.

Click on Analytics icon on the left to switch to Analytics panel. Choose and drag & drop the selected function to the canvas.

When you release your left-mouse button, selected function displays.

With Properties, you can set detailed properties of an analytical function. As you can see, to apply these analytics to the visualization, there is no coding required. Just Drag drop and set properties if needed.

Let’s look at individual Analytics now.

Reference Line:

Reference Line is a line that represents a simple descriptive statistic like Minimum, Maximum, Average and Median. There is another option to present a constant line on the graph as well.

Additionally, users have an option to choose a method of how they would like the reference line to be displayed – line or band. When Band is selected, the minimum and maximum reference lines for the band must be selected.

Trend Line:

Trend line is a line that fits a linear, exponential or polynomial model and returns the fitted values which are then displayed on the visualization.

There are 3 different methods of how data fit actual values and display a trend:

– Linear,

– Exponential and

– Polynomial.

The example above is showing a polynomial trend line with a degree of 3. The grey area around the trend line is showing a confidence interval of 95%. You have an option to choose among 90%, 95% and 99% confidence interval. You can also decide to switch display confidence interval off.

Forecast:

With Forecast, you can predict values for the next n future periods. Number of n next periods can be set as required.

Clusters:

With Clusters Analytical function you can collect a set of records into groups based on one or more input expressions. There are two algorithms that can be used for clustering, K-Means and Hierarchical Clustering.

With Properties, you can set the number of clusters, i.e. marketing segments. On the chart above it is clearly seen how clusters are created and what are the rules for it.

 Outliers:

 Outlier is a function that classifies a record as an outlier based on one or more input expressions using K-Means or Hierarchical Clustering  Outlier detection algorithms.

Sometimes it is required to identify outliers and exclude them from the analysis. For example, if you are working on some classification algorithm which would predict withdraws from the ATM machine.

CONCLUSION:

As you can see, Data Visualization Desktop has built-in Analytics on “click” that can be deployed without any coding, just by dragging & dropping.