Big data as art

The The Of Of Analytics project from Teradata looks rather unusual. The idea of ​​the project is to explain research based on big data in the form of artistic images to a wide audience. Want to find out what the detection of bank fraud and terrorist threats looks like or a comparison of single malt whiskey? Under the cut, some of the 20 studies depicted as paintings.



Teradata experts in the field of big data analysis managed to explain complex things in a simple language, expressing the most complex mathematical relationships in the language of art. Instead of graphs and numbers, each case of working with bid data is represented by an abstract picture resembling associative graphics. The Art of Analytics project consists of 20 paintings. No matter how simple they look, this is the result of really painstaking work of world scientists in the field of big data.

Epicenter of the storm (Eye of The Storm)

In my opinion, the most interesting project, the visualization of which at the same time is similar to the eye. This is big data analytics during a twitter storm (a huge number of tweets and retweets immediately appear on a tweet) about the death of Margaret Thatcher.

• Red dots: haters users .
• White Dots: Iron Lady Politics
• Yellow: opposition.
• All the lines between them are informational communications (who answered whom and to whom).

The user publishes a tweet - a new point appears in the picture. Thus, you can find out who was the "starting point" of a post, understand the reasons and identify the perpetrators. Such mechanisms can at least find the instigators of terrorist acts, or even prevent them. Such an analysis of activity on social networks would be very useful, for example, for Arab governments shaken by the 2010 Arab Spring revolutions - after all, the opposition prepared and coordinated the main protests on Facebook and Twitter.
Data was downloaded using the twitterR package and collected for one week, then it was uploaded to Teradata Aster for further processing. Then Aster Lens was used to create the graph.

Financial Fountains

This cute bouquet of Lankaran acacia is all the movement of money between different companies that are marked in different colors.

• Points: companies.
• Lines: number of transactions: the more lines, the more successful the company.

A large company will look at large flowers and cooperate with them, protecting itself from risks, and frail flowers - risky enterprises - will attract the attention of various support funds, without affecting the entire market and without jeopardizing the economy as a whole. One use is for the tax police to use such analytics to initiate a company audit for illegal financial transactions.
Transaction data uploaded for processing: more than 60 million records of 670 thousand companies. The graph has 3883 points and 3943 lines.

Stargate

Anonymized visualization is part of an analytical project conducted in Taiwan for one mobile operator seeking to stop malicious Internet attacks against itself.
"Stargate" is aimed at identifying and stopping - Advanced Persistent Threat (APT) attacks , which are hidden and lengthy processes of hacking computers and systems aimed at a specific organization or an entire country for political or business reasons.

• Dots: Web pages opened by the user.
• Ribs: paths to other pages.
• Line width: characterizes the length of time spent on the site, and the size of the nodes represents an indicator calculated on the basis of 128 characteristics of the web page (category, reputation rating, frequency of visits, content, TTL- prescribed lifetime of the forwarded packet).
• Larger nodes in the center of the picture: they are high-risk sites that lead to malicious behavior, on which it makes sense to increase the security policy, as an example of a WAL (Web application firewall).

The analysis covers the log data of 2,300 employees for the period of operation of the system for one year. This analysis uses nPath, cluster and text analytics.

Single Salt Sampler

Can analytics radically simplify all the subtleties and nuances accumulated over the 700-year history of Scotch whiskey?
This type of analytics is used in food science. It can be combined with market share and segment profile data to create new flavors or to reposition existing brands in new segments.
For whiskey fans, he also offers an exciting way to explore this extremely complex subject area. If you like the taste profile of one brand, you can try others similar to it, or discover completely different types of flavor combinations.

• Points: brands of single malt whiskey.
• Ribs: the degree of similarity between the whiskeys - the thicker and darker the line, the more similar the taste characteristics.

Visualization consists of 86 single-malt brands, grouped by 12 taste characteristics (sweetness, honey notes, nutty flavor, etc.)

Calling Circles

Where does the information about who you call, when, how long you talk, who receives your SMS and MMS get to? These data enter the database of the mobile operator, or any organization that has access to them. This information is a classic example of "big data": its volume is enormous.

• Dots: phone numbers.
• Ribs: calls where long - long calls, short - short.

The line connects 2 points, that is, the one who has access to such information knows not only that you have a mobile phone, but also how often you use it, who you call and who calls you. Why is it necessary? Such analytics is very useful for mobile operators, which in this way can adjust tariffs and programs to the needs of subscribers and track the habits and preferences of users.

Basic analysis tools:

• Teradata Aster - data processing.
• Aster Lens / Quartal Super Graphics / Gephi - data visualization.

Project Links:

1. The official page of the Art of Analytics project , studies will be published here.
2. Another project page .