Hi. Last week, you could identify the requirements of a typical alarm system, and you have proposed a columnar database to have security, concurrency and excellent performance on analytical queries on a structured data. Are you ready for the next use-case? Remember that your proposal must conform to what it is requested. No more no less. This style would be use-case of this week. A chain of bookstores requires to develop a data-intensive application that can contain book sales transactions. Some of these transactions have been reported by customers as unknown or fraudulent. There are millions of transactions per day and a customer come by on different bookstores on the same day. In order to implement a database, suppose there are 16 bookstores, 10 customers. Each customer has bought on at least three different stores. The database contains a history of 60 transactions; 16 of them are fraudulent, are having mark as complaint, the rest are normal transaction or no-complained. Remember that the last type of no-sequel database that we mentioned corresponds to graph oriented databases while a graph is represented as a set of nodes or entities interconnected by edge or relationships. Graphs give importance not only to the data but to the relations between them. Relationships can also have attributes and direct queries can be made to relationships rather than to the nodes. Being stored in this way it is much more efficient to navigate between relationships than in a relational model. Obviously these type of databases are only useful when information can be easily represented as a network. As a summary on Neo4j we can say that it is a graph database with full Acid conformance including long-lasting data. With nodes such as relationships that can have meta-data. It's query language is called Cipher, and it is based on integrated pattern-matching. Neo4j is best used for graphic style data, rich or complex interconnected. Companies loose billions of dollars every year to credit card fraud. Credit card data can be stolen by criminals using a variety of methods. For instance Bluetooth-enabled data scheming devices can be placed on the card reader. The data might be stolen in mass breach by hackers of a large retailer. Sometimes the thief is simply the clerk at the checkout line at the bookstore while the big Kim's card is sweeped through a small device also surreptitiously jotted down. The center data-intensive application that can help to fraud detection with simple queries like: A, obtain all customers that have complained about fraudulent transactions or books they have not actually bought, show which customers and which bookshops are involved in the fraud cases. B, the bookstore chain not only want the illegitimate transactions but also the transactions happening before the theft. C, find the common bookshop in all this fraudulent transactions. As we have learned in course two, how would be convenient to start analyzing such amounts of data to get the best business advantages? Well this is where predictive analytics, data mining, machine learning and decision management come into play. Data mining is a process of discovering patterns in large data sets involving methods at the intersection of machine learning, statistics and database systems. Data mining is used to improve all lab analyses. Data mining is also used within the process of knowledge data discovery. Remember that predictive data mining tasks come up with a model from unavailable dataset to predict unknown or future values of another dataset. For instance a medical practitioner trying to diagnose a disease based on the medical test results of a patient and descriptive did data mining thus find data describing patterns and come up with new significant information from the available data set. What solution would you bring to the company? The students would submit its proposal and justification of the following elements: One, typical architecture according to the type of information system. Two, kind of database to implement and it's design. Three, answers to the queries required. Four, how does it support scalability. Five, how does it support maintainability. Six, how does it support security and reliability. Well, that's all for today. Good luck.
