 Hello. Myself Dr. R. V. R. Giddy from Computer Science and Engineering Department of Walsh and Institute of Technology Sholapur. The today's video topic is on data processing and mining. The topic name is data models. The learning outcome of today's video of this video are student will be able to compare various data models they will be able to integrate OLAB for data mining and knowledge discovery activities. The content of this video we will see the comparison first of all of operational database and data warehouse operational database like the database we use normally that is SQL access, no SQL, ingress all those things and the multidimensional data model then we will see the conceptual modeling of data warehouse. First of all let us focus on difference between operational database and data warehouse. In operational database that is what we use SQL or no SQL or access it involves day to day processing whereas in data warehouse it involves historical processing of information. In data base it is OLTP that is online transactional processing systems are used by it is used by clerks or database administrator or database professionals. Whereas data warehouse it is actually online analytical processing the systems are used by knowledge workers such as the executives of the organization managers and the analyst of that particular domain. The third point that is operational database it is used to run the business whereas it focuses on information what is the information coming out. Then in databases it is used for online transactional processing and in data warehouse it is used for online analytical processing. These are the differences between operational database and a data warehouse. Let us take the next point that is multidimensional data models. A data warehouse is based on multidimensional data model which views data in the form of data cube. This means data is represented in the form of cube instead of in one-dimensional instead of representing in n-dimensional cubes are used to represent the data. Then a data cube such as cells means here we are considering one example if you consider the cells of a particular organization. Then in the cells we have item what are the different items and at what time the cells has been occurred in which place, location all those things can be considered. Now representing a particular data in one-dimensional you can represent either item or either time or total number of purchase. But in multidimensional you can represent all these attributes at same time in a cube. Now fact table contains measures such as dollar, sword and keys to each of the real-time dimensional tables. In data warehouse literature an n-dimensional cube is called a base cubite. The top most cubite is called a zero dimensional cubite which holds the highest level of summarization is called the effects of the cube. And the lattice of the cubite forms a data cube. Now this is a cube multidimensional data model in which we have shown the time, item, location and supplier. Means the fields of the fields of cells which are the items they have been sold at what time or in which month or in which quarter of the year and at which location and who was the supplier. Now this is effects of the cubite that is at zero dimensional. At one dimensional these fields are shown that is time, item, location and supplier. At two dimensional now we consider time and item combining these two the time and item is shown over here at two dimensional. And if you combine time and location then that particular those two attributes of time and location are shown in the two dimensional. Similarly if you consider time and supplier and location and supplier those points are shown over here as item location and location supplier. This is at two dimensional. At three dimensional if you go down then now we consider time, location and item location. Combining these two fields we get a data, a common attribute will be eliminated. Like here the common is location. So it will be time, location and item will be displayed at the three dimensional level. And then combining these two item, location and location supplier. Here again the location is combined common attribute. So item, location and supplier will be shown over here. Similarly time, item and supplier is shown when time, location and location supplier is combined. At the fourth level if you combine all those fields then we get common items, common attributes are time, item, location and supplier. We get the information of all these data. This is how we can represent a data in multi-dimensional. We can find out the sales data of a particular item at what time, at which location and who was the supplier. All those data can be shown in a different levels. Now let us think, I will give the question what is the dimensional modeling? Think on this particular question, take a pause over here and try to give the answer. Now dimensional modeling a different way to view and cross examine data in a database. The second point, a dimension is a collection of logically related attributes and is viewed as an axis for modeling the data. For example, the time dimension can be divided into decade, year, month, week, day, hour, minute or second. So one particular time or field can be divided into these many attributes. Similarly item can be divided into name of the item, code of the item, number of item and so on, cost of the item. So these are the attributes of that particular field. Now conceptual modeling of data warehouse. Let us see. Now modeling data warehouse, there are three types of modeling. The first one is star schema. It is a fact table in the middle connected to a set of dimensional tables. Then the snowflake schema, refinement of star schema where some dimensional hierarchy is normalized into a set of smaller dimensional table forming a shape similar to snowflake. We will look at the diagram and I will explain with the diagram. And the third is fact constellation. Let us look at the example of star schema. Here at the center, all the attributes are, all the fields are mentioned. Now what star schema says that time key is again further divided or subdivided into time key, day, day of the week, month, quarter and year. Then the branch key, it can be again further divided into branch key, branch name and branch type. And item key can be divided further. Item key, item name, brand, type, supply range, similarly location. So this shows that the fields are further divided into attributes. Now snowflake, whatever the fields they are divided in the first level, they are again divided into second level. Like item key is divided into item key, item name, brand name and supplier key. That is again further divided into supplier key and supplier type. Similarly, time key is divided, there is no further distribution. Location is distributed of that city key is considered and that is further divided into city key, name of the city and name of the country. Then the fact constellation, now this will be reverse from small part to the higher part. That is item key is again divided into item key and item name and so on. And here from this part also item key is divided into means from both the sides this can be shown. The reference for this particular video was data mining. These two books were used. I hope you have understood. Thank you.