One of the Oracle Public Cloud Services I have seen mouthwatering demos with but have not actually tried out myself is Oracle Data Visualization Cloud. I had several triggers to at last give it a try – and I am glad I did. In this article a brief report of my first experiences with this cloud service that aims to provide the business user – aka citizen data scientist – with the means that explore data and come up with insights and meaningful visualizations that can easily be shared across the team, department or enterprise.
I got myself a 30-day trial to the cloud service, uploaded a simple Excel document with a publicly available dataset on the countries of the world and started to play around. It turned out to be quite simple – and a lot of fun – to come up with some interesting findings and visualizations. No technical skills required – certainly not any beyond an average Excel user.
- get a trial account – it took about 5 hours from my initial request for a trial to the moment the trial was confirmed and the service had been provisioned
- enter the Data Viz CS and add a new Data Source (from the Excel file with country data)
- do a little data preparation (assign attributes to be used as measures) on the data source
- create a new project; try out Smart Insights for initial data exploration, to get a feel for attributes
- create some visualizations – get a feel for the data and for what DVCS can do ; many different types of visualizations, various options for filtering, embellishing, highlighting; many dimensions to be included in a visualization
- try a narrative – a dossier with multiple visualizations, to tell my story
In a few hours, you get a very good feel for what can be done.
Create Countries as my new data source
First steps: download countries.csv from https://www.laenderdaten.info/downloads/ . Create Excel workbook from this data. Note: I first tried to upload the raw csv file format, but that ended with an internal error.
The data that was imported is shown. Now is a good moment to set the record straight – any meta data defined at this point is inherited in projects and visualizations that use this data. For example: if we want to calculate with attributes and use them as values – for the size of bubbles and stacks and to plot a line – we have to identify those attributes as measures.
Once the data source is set up – we can create our first project based on that data source by simply clicking on it. Note: the PROPOSED_ACTS and ACT_ALBUMS data sources are based on database table in an Oracle DBaaS instance to which I have first created a connection (simply with host, port, service name and username & password).
My First Project – Data Preparation & Initial Exploration
Here is the Data Preparation page in the project. We can review the data, see what is there, modify the definition of attributes, prescribe a certain treatment (conversion) of data, add (derived) attributes etc.
If we click on the Visuals icon in the upper right hand corner, we get a first stab at visualization of some of the data in this data source. Out of the box – just based on how DVCS interprets the data and the various attributes:
For example the number of countries per continent. Note how we can select different measures from the dropdownlist – for example area:
This tells us that Asia is the largest continent in landmass, followed by Africa, Russia is the largest country and the size of all countries using a currency called Dollar put together is the largest, with Rubel using countries (probably just one) as a runner up. Note: all of this – out of the box. I am 10 minutes into my exploration of DVCS!
First Visualizations – Let’s try a Few Things
Go to the second tab – Visualize.
Drag an attribute – or multiple attributes – to the canvas.
The default visualization is a pivot table that presents the data for the selected attribute. Drag two more attributes to the canvas:
The result is a matrix of data – with the measure (area) in the cells:
In order to prepare the visualization for presentation and sharing, we can do several things – such as removing columns or rows of data that is not relevant, setting a color to highlight a cell:
Define range filters on selected attributes – for example filter on countries with at least a 45M population or
When the filter has been set, the matrix adapts:
We can try out different styles of visualization – what about a map?
DVCS recognizes the names of continents and countries as geographical indications and can represent them on a map, using color for total area. Let’s remove continent from the Category dimensions, and let’s set bubble size for population:
If we are interested in population density, we can add a calculated value:
Some more examples:
Select countries by size per continent – in a horizontal stack chart:
and a treemap – with population added in as additional attribute represented through color:
Any visualization we like and want to include in our final narrative can be saved as an insight:
On the Narrate tab – we can include these insights in a meaningful order to tell our story through the data visualizations. Also see Building Stories.
Oracle Data Visualization Cloud Service: https://cloud.oracle.com/en_US/data-visualization (at $75.00 / Named User / Month with a minimum of 5 users a fairly friendly priced offering)
The country data is downloaded from https://www.laenderdaten.info/downloads/
Documentation on Oracle Data Visualization Cloud Service: http://docs.oracle.com/en/cloud/paas/data-visualization-cloud/bidvc/getting-started-oracle-data-visualization.html
Documentation on Building Stories: http://docs.oracle.com/en/cloud/paas/data-visualization-cloud/bidvc/building-stories.html#GUID-9D6282AA-C7B7-4F7E-9B9E-873EF8F1FB5D