Data visualization

One of experts in the field of data visualization Edward Tufte told once: "The world is difficult, dynamic, multidimensional, and paper static and flat. How can we transfer all richness of visual representation about this world on this plain?". In fact, the developed visualization tools are obligatory components of both industrial BI platforms, and small analytical solutions for end users today.

Data visualization, according to experts of Forrester The Forrester Wave^[1], is necessary for the company when:

• It is impossible to see a pattern, without resorting to data visualization. All because traditional reports, lines, columns and tables do not give understanding of an overall picture in general that can lead to incorrect interpretation of data.

• When considerable data arrays cannot be placed on one screen. How the font was small, hundreds and thousands of lines of data on one screen will not be informative in comparison with visual submission of the same information at all. In the^[2] Edward Taft gives an example of display of 21 thousand values of data on the USA map which was located on one screen.

• It is impossible to transfer depth of data by means of information provided on one screen. For example, in pharmaceutics in the analysis of properties of these or those medicines hundreds of different psychophysical parameters of the person which are not possible for covering in a linear type are used. However medical experts appropriate to each parameter multi-colored pixels and place them in dynamic models where any section if necessary can be increased and analyzed in details.

Example of visualization of Symphony No. 5 of Ludwig Van Beethoven

Many companies for long years effectively use such simplest visualization tools of data as charts and histograms, further development of technologies of visual representation of information allowed to make work with them of more dynamic and to process data actually in real time. The developed business graphics includes different deshborda and charts which can automatically be updated according to the set settings.

Example of visualization: communications between subjects of reports of the different countries within the World Economic Forum

Example of data view in the form of a deshbord

Among variety of separate tools and systems with functions of data visualization of Forrester Research also selects separately platforms of the visualization advanced date or ADV (advanced data visualization). Such platforms have more developed means and provide more complex information representation. There are their main defining differences:

• Dynamic content. It means that visual representation of information changes in process of an update of data in base. Static visualization which can be created in the majority of office programs does not give such opportunity.

• Visual requests. It is an opportunity to request and perezaprashivat the diagrams given by simple manipulation by parts or charts (clicking buttons, for example, it is possible to unroll details) or special visual components (drop-down lists, a vladok and other).

• Multidata. Typical static visualization tools cannot connect and display among themselves more than several parameters of data. In advanced platforms of visualization, on the contrary, the multiple data of different types firmly tied with each other so when updating some part information automatically changes on all other panels are considered.

• The animated visualization. Some parameters, for example, time, can have hundreds and hundreds of values so tracking of an overall picture for each parameter separately can become rather labor-consuming task. In such cases the animation allowing to start transition from parameter to parameter in the automatic mode comes to the rescue.

• Personalisation. What intuitively is also necessary for one user can not suit another at all. Besides, the different access levels to corporate data are provided in most the companies for data protection and risk minimization. The ADV platforms allow to apply personal settings and preferences automatically.

• Use of notifications. Powerful ADV platforms can give a signal when is not able to process data bulks correctly. Or, for example, the signal is given while some critical value is found, and the notification can move including using e-mail and text messages.

Market leaders of platforms of an advanced vizualiation of data

Forrester Research, 2012

According to Forrester Research, such companies as Tableau, IBM, Information Builders, SAS, SAP, Tibco, and Oracle treat leaders in the market of ADV platforms. Solutions of these vendors give the considerable opportunities for visualization well balanced for creation of different requests.

Strong players in this market are also Microsoft, MicroStrategy, Actuate Software, Qlik (QlikTech), SpagoBI and Panorama Software. An alternative mega to vendors are also Jaspersoft and Pentaho whose solutions can be used on a subscription.

As the expert Georgy Naneishvili, the development director of a partner network of Qlik told TAdviser, the Russian companies as visualization tools prefer traditionally tables – simple and summary.

"It is good to show beautiful diagrams on the presentations, but in real work for analysts crucially instantly to work with huge information volume, dynamically imposing filters and receiving reports in a tabular form. We will not leave from technology of data analysis by a stare method in any way. And the acute shortage of the qualified infograf is felt", - he explained.

So to afford really effective solution which one look also historical trends will allow to assess current situation, to carry out contrastive analysis, the geoanalysis – are able to afford not so many Russian companies, he believes.

And case not the fact that there are no similar solutions: on the contrary, all modern BI tools are focused on it, and especially advanced allow to do it even to ordinary users. The expert believes that for the present users have no understanding and sufficient qualification.

the Forecast for trends in medical visualization

In December, 2017 the CEO of Ambra Health Morris Panner submitted the forecast for technology trends which, according to him, will change medical visualization in 2018. Read more here.

Visualization of Big Data

The basic value for their interpretation has an evident idea of analysis results of Big Data. It is no secret that perception of the person is limited, and scientists continue to conduct researches in the field of improvement of modern methods of data view in the form of images, charts or animation. It would seem, nothing new can already be thought up here, but actually it not so. By way of illustration we provide several progressive methods of visualization which rather recently gained distribution.

• Tag cloud

A certain weighting coefficient which correlates with a type size is appropriated to each element in a cloud of the tag. In case of the analysis of the text the value of a weighting coefficient directly depends on the frequency of the use (citing) of a certain word or a phrase. Allows the reader to gain in a short time an impression about the key moments of as much as necessary big text or a set of texts.

• Klastergramma

The visualization method which is used at cluster analysis. Shows as separate set members of data correspond to clusters in process of change of their quantity. The choice of optimal quantity of clusters – an important component of cluster analysis.

• Historical flow

Helps to monitor evolution of the document on which creation at the same time large number of authors works. In particular, it is a typical situation for wiki services and the website of TAdviser including. On a horizontal axis time, on vertical – a deposit of each of coauthors, i.e. volume of the entered text is postponed. A certain color on the chart is appropriated to each unique author. The provided chart – an analysis result for the word "Islam" in Wikipedia. It is well visible how the activity of authors increased eventually.

• Space flow

This chart allows to monitor space distribution of information. The chart given as an example is constructed using New York Talk Exchange service. It visualizes intensity of exchange of IP traffic between New York and other cities of the world. The line is brighter – the more data are transferred for unit of time. It it is easy, it is possible to select the regions, the closest to New York in the context of information exchange.

Processing of operational databases and multidimensional storages

• Transaction heavy-duty databases
• Relational and multidimensional storages
• Seamless accesses to API systems
• Work with connectors
• Processing of flat files (CSV, Excel, XML, etc.)

Creation of analytical models in memory

• Classical BI solutions
• Hybrid approach (Exalytics, HANA, Teradata, etc.)
• In-memory of a system

• • Essentially other approach
• • Not relational databases
• • Compression
• • Access rate
• • Indexation and search
• • Visualization tools

Typical problems of visualization and calculations "on the fly"

• Attempt to create the interface overloaded with elements
• Redundancy of data (millions of lines in tabular and graphic elements)
• Dynamic measure calculation
• Difficult calculations
• Conditional formatting
• The non-optimal expressions leading to complete scanning of tables
• Payment under large volumes of primary data

The tools combining ETL tools and visualization tools

Tasks

• SQL requests
• Work with stored procedures
• Work with flat files as with tables of base
• Scripts of conversion, enrichment and data scrubbing
• Work with functions
• Formation of new measurements and values
• Preparation of relational communications between entities

Predictive and statistical analytics

• Data processing, formation of local forecast model, implementation of mathematical algorithms
• Data transmission and use of external tools
• Work with DBMS and use of functions of processing of data bulks
• Multithreaded processing
• Comparative analytics
• What-If analysis and parametrization of reports

See Also

Business Intelligence, BI (world market)

Trends of development of the world market of BI

Business Intelligence (Russian market)

CPM (world market)

Big Data world market

Self-Service BI

Predicative analytics (predictive, forecast, predictive) Predictive analytics

Cloud/SaaS BI

Notes