Data visualisation consists of graphically representing a set of information, as opposed to a textual representation.
Today, it's clear that the field of visualisation goes well beyond this restrictive definition:
This means it is essential to apply transformations on any dataset before tackling its visual representation. Although these two aspects are inextricably linked, a distinction needs to be made between them as they involve very different skills and methods.
This consideration brings into the debate the issue of the database on which the visualisations are based, and this second point will prove as critical as the data representation functions.
It's clear that the choice of processing and the tools used will be radically different:
The tools can be difficult to choose, both because of the number of vendors on the market and the technical aspects of the subject, which all too often take precedence over organisational and governance considerations.
Aside from the mere economic equation, the selection criteria to contrast with the aforementioned challenges are listed below:
The vendor's own brief words to position is product.
Some tools provide extremely sophisticated options, although they naturally require suitable skills. This capability is often at odds with learning to use the tool intuitively. Conversely, tools with limited options have the advantage of uptake by a larger user population, without requiring a significant training effort.
What capabilities does the tool have to convert, consolidate and remodel different sources to integrate them into a single report? The tool can also provide features that guarantee the consistency of the information presented.
What features are provided that allow users to collaborate on the same report (comments, sharing, etc.)?
A visualisation can accelerate decision-making, although it still needs to be accessible. Even disregarding the restriction of use within a company network, what access is possible without requiring a thick client? In a standard browser, a dedicated application? Is the tool responsive? Etc.
The data world has been radically transformed in recent years. Even only considering the aspects of volume and performance, it's now completely conceivable,or in some cases advisable, to give a user access to very large datasets. A vendor's ability to offer tools that respond to these new challenges must also taken into account.
This introduction is an extract from the chapter dedicated to Smart Data & Dataviz in the Yearbook 2019 (M13h) ; click below to download the long version:
Resources on Smart Data & Dataviz
Video of the @Berlin2019 conference
This new topic will present solutions for preparing, exploiting and making intelligible a large amount of data.