Hi. Last week we learned how data mining allows descriptive and predictive analysis of data. This week we will learn some data types according to the structure of data and how to integrate and store and analyze unstructured data. We have already learned that there are huge amounts of data that people need to analyze and the data comes from different disparate data sources. Well, as you may know, there are structured, semi-structured, and unstructured data. We have already learned that in traditional relational databases, we store structured data. Now we'll see how to integrate and store and analyze any type of data. Structured data can be stored in a relational database. They have a structure that corresponds to their schema with a column name, data type et cetera. The simplest way to manage information is structured data, but it represents only five to 10 percent of all informatics data. Semi-structured data is form of structured data that does not confer with the formal structure but nonetheless, contains tags or other markers to separate semantic elements and enforce hierarchies of records and fields within the data. Examples of semi-structured data are Comma-Separated Values but Extensible Markup Language or XML and JavaScript Object Notation or JSON documents are semi-structured documents. NoSQL databases are considered as semi structured because they provide a flexible schema. In the case of non-structured data, all the remaining data having no structure at all falls into this category and it represents the 90 percent of total information. Examples could be censored data, social media streams, images, videos, mobile data et cetera. Starting from the emergence of the Internet and therefore of social networks, data that flows through the organization is mostly semi-structured or unstructured. This has changed the paradigm of data management that was initially through database manager who's gone to request the structure for storage. From these, programmers have a reason for the integration, cleaning, and analysis of different sources of information. Now we can see the different data sources and how they need to be integrated to be analyzed. Nowadays, business strategies, marketing, and the way of establishing customer relationships have dramatically changed. Currently, you cannot conceive promoting or selling a product without the use of social networks. Emails, tweets, online sales are the main means to approach your future customers and keep existing ones. In the context of the problem integration, remember the operational applications. Those are generated operation and their daily business are part of the information sources that we wish to analyze. Originally, they were analyzed through a data warehouse that requires section and transformation of OLTP data and loaded into these data warehouse in an OLAP system. Now both OLTP and OLAP applications are sources of information that you want to analyze with algorithms and increasingly sophisticated tools do not only describe events but predict and prescribe future events. Let's remember the OLTP characteristics. In the case of our lab applications, useful data were stored on a database manager because they double as a structure. Nowadays, analytical applications require to integrate OLAP, OLTP and social network data. Let's remember the OLAP characteristics and the OLAP architecture. In the next video, we will see how to manage unstructured data. See you soon.