Large collections of data (Big Data) offer companies the chance to supply customers with highly relevant and specific data. censhare has enhanced this feature in Version 4.8. context information, such as areas of interest or recommendations for a topic, can be given a weighting.
Extracting relevant information from the deluge of data
The current catch phrase "Big Data" refers to the challenge for companies of storing fast growing quantities of data and using them for their business. With Context Aware Computing you can offer your customers just the part of the information which is relevant to them in their situation. censhare can store and evaluate data for Context Aware Computing.
The more information a user supplies about their current situation (context) and their areas of interest, the better the selection can be. These data are then stored and interlinked. This creates a network which grows with every additional item of information on the user. This digital profile then forms the basis for offering appropriate contents, services, products, etc., automatically to users. In Version 4.8, censhare extends the context information with a weighting of the data. This improves the ability to describe and act on the meanings and interconnections among the data even further.
Such information can, for example, make it much simpler to offer users hits which are relevant to them. A company could provide the descriptions of its books with keywords. Since not all expressions are equally relevant, they receive a weighting in addition. A travel story about the flight of a young boy through Ireland in the 19th century might contain the following tags, among others: novel (100 percent), adventure (90 percent) and travel (70 percent). An object or asset in censhare can be provided with keywords which receive a weighting in percent. If a user is looking for novels, the book appears much higher up the hit list than if travel was the criterion. Besides keywords, objects in censhare can also be flexibly extended with other properties. The book mentioned above could also receive the target groups youth (100 percent) and young adults (50 percent).
Besides improving the buying experience throughout the web, Context-Aware Computing offers features like the supply of information dependent on media channel and location. When users are on the move with their smartphones, they are interested in other things than when they are at home surfing the Internet. censhare provides the basis for capturing the context alongside the actual information.
This could be sensor information such as time or location, attributes such as age or company, relationships to other people or historical data. The latter could include purchase actions or articles read. censhare stores this information as objects with properties and connects them with links, for example User and Interests. The links and properties can also be ascribed a relevance.
Objects, properties, links and relevance thereby combine to form a comprehensive data model in censhare to represent reality with its interconnections. Both information and users can be described comprehensively in their particular context. With this model, companies can extract the greatest possible benefits from the data available.
Storing and processing large quantities of unstructured data or information
Development of context-dependent applications (Context-Aware Computing)
Comprehensive data model with objects, properties, links and relevance in order to store and use data as effectively as possible
Definition of objects such as user, company or products with their environment (context)
Selection of appropriate data or content in applications dependent on the context of the user and the application
Representation of complex and unstructured data with censhare's data model, as they occur in Big Data and Context-Aware Computing
Definition of the relevance of an item of Information such as keywords, target groups, geo-coordinates or categories for an object
Weighting of the results of a search for particular properties or links with the help of the associated relevance
Linking different database approaches (NoSQL) for working with large quantities of unstructured data
Combination of the advantages of relational and NoSQL databases without their disadvantages
Scalable architecture for application to large volumes of data and a large number of users
Development and provision of content applications which take users, their interests, the language, the device or the location into account
Different presentation and volume of content, dependent on device used, such as laptop, tablet or smartphone
Personalization of the contents for different target groups based on user data such as interests, profession or age
Everything for storing the real world
Relational databases are not suitable for storing large quantities of unstructured data (Big Data) as required for Context-Aware Computing. Consequently, censhare uses a graph model where objects such as persons, sensor data or companies are linked to each other by relationships. This data model is further refined with the relevance property. In response to a search, the results with a higher relevance are displayed higher up the list. The graph model is also one of the approaches used by NoSQL (Not-only-SQL) databases. The topic of Big Data has increased their significance as an alternative to relational databases.
Documents form a part of the unstructured information which occurs in Big Data. XML lends itself very well to describing the information in these documents. censhare can also store and process this kind of data appropriately. It uses XML for the description of documents and the internal structures. XSLT (XSL Transformation) can be employed to change this information automatically. Document databases, which also belong to the NoSQL databases, are similarly based on XML.
Among the challenges of Big Data there is the need to sort through the large amount of data available in almost real time and to deliver answers. With the help of a database developed in-house, censhare can query its data sets in seconds. In this context, the database applies a Key-Value approach where each line consists solely of a key and the associated value. The Key-Value approach is also applied in NoSQL databases.
NoSQL Databases are still a relatively young technology. This is why they are not yet as mature as relational databases in many areas. censhare has succeeded in combining the best of both worlds. For work with large quantities of data (logical database view) it behaves like a NoSQL database. Physical storage, however, takes place in a relational database with all the advantages such as transactional security or mature backup and archiving tools. But censhare does not have the problems that arise from relational databases' logical data models when they are applied to Big Data. In this context, there is a similarity to NoSQL databases.
Very close to the real world
In censhare's data model, objects are represented by assets. Complicated objects, such as a country as a travel destination, can be built up from assets. For example, Ireland could be composed of regions. These may possess interesting sights or towns. Towns, in turn, also have interesting sights. The various assets are connected with one another by directed links which may also be followed. The most varied information such as text, images, audio or videos may be associated with an asset.
A user reads a contribution on an interesting sight in a particular town. Next, the individual would like to learn something about further sights in the town. From the current article, censhare finds the object for the associated town. censhare can then follow the links to all the sights which are connected with the town, and offer these to the user. The advantage of this approach to searching is that the response time does not depend on the total data volume, but only on the size of the results. In this case, it is the number of sights for the current town and not the total of interesting sights.
censhare's data model is very flexible and can easily be extended without any need to change the existing structure. censhare has extended the data model with a further optional property, namely that of relevance. In the Admin Client this can be activated very simply with a tick in the appropriate dialog for properties or links. In the above example it could represent the popularity of a town or a sight. If the user searches for further sights in a town, censhare can sort the results by their relevance. In this case the user does not receive the results in alphabetical order, but sorted by the popularity of the interesting sights. In this way, the most interesting places appear right at the top of the list.
After activation, users can enter a value against the associated property in the dialog for the properties (metadata) of an object (asset). The default value is 100 percent.
With this data model censhare is very well positioned for applications in the area of Big Data or Context-Aware Computing. Unstructured data, which are often present in Big Data, can be represented well.