 Hello everyone, I am Gannapai Valikar, work in the Department of Computer Science and Engineering, Vulture Institute of Technology, Sholapur. Today I am going to deliver a lecture on data acquiring, organization, processing and analytics in IoT. As part of learning outcome at the end of this session students will be able to describe data acquiring process, data organization process, data processing and data analytics in IoT. Before discussing different data architecture and Berkley stack architecture, let's we will discuss basics of how to acquire the data and how to process the data and then how to store that process data. And this is the standard model where things are acquiring the data, then processing the data, then process data will stored in a local or a another storing devices and then how it will be communicating with a third party. The sensors, activators, they will sense the information, then it will send it to the second data acquisition model. Then acquisition model will convert that acquired physical signal into a digital signals. Then data processing model will show that a converted digital signals and process that digital signals and perform the analytics, then store it either locally and other rates computing. So once it is analyzed and stored and that will be communicated to the user via communication model. And this is how components are interacting each other while acquiring the data and storing the data. Security devices will sense the data and produces the report and send that report to the second component that is a processing component. Then the second component will receive the raw information and process it and stored in the storage media. So once it is stored locally or other rates computing devices and that data will get analyzed in order to extract the knowledge and this extraction of the knowledge might be iterative and many times it may be happens until and unless get the final knowledge from the stored data. So after getting the knowledge or analyzing the data, the feedback about that data will send back to the sensor devices. So now we will discuss about so how data being acquired then how it will be stored. There are four steps in the data acquiring and storing of a data. First one is data generation, second one is data acquisition, third one is data validation, the last one is data storage. Now we will discuss one by one, what do you mean by data generation, what do you mean by data acquisition, what do you mean by data validation and data storage. Data generation is a process of generating the data through passive devices, data or active devices, event data or device real-time data and depending upon the type of the devices generated the data will be categorized as passive device data, active device data, event data, device real-time data and event driven device data. So once the data is generated using different data generation devices then we have a different type of data. So then once the data is available data is acquired and that means the acquiring data from IOTR a M2M device. So once the data is acquired so then we need to be validate that data for the complete, correct and consistent. Data acquired from the devices does not mean that data are correct, meaningful or consistent. So that need to be done in the validation process. So once the validation is done to ensure acquired data is complete, consistent and correct then we have to store that validated data in the repository. So once the data is acquired then we need to organize data in the IOTR. The organization of data will be done either through RDBMS or distributed database or query processing. Let us discuss how data being organized through RDBMS, through DDB and query processing. A relational database is a collection of data into multiple tables which relate to each other through special fields called key. Example, MySQL, PostgreSQL, Oracle, PL, SQL. The distributed database is a collection of logically interrelated databases over a computer network. The distributed RDBMS manages distributed databases. The query processing means using processes and getting the result of the query made from the database. So now we will discuss data analytics in IOT. Data analytics in IOT is the application of data analysis tools and procedures to realize value from the use volumes of data generated by connected IOT devices. Data analytics is the process of examining data sets in order to draw conclusions about the information they contain. The analytics pages in IOT are descriptive analytics, redictor analytics and prospective analysis. Let us we will discuss one by one. The descriptive analytics enables deriving the additional value of visualizations and reports. The predictive analytics is advanced analytics which enables extraction of new facts and knowledge and then predicts or forecasts. The prospective analytics enables derivation of the additional value and undertake better decisions for a new options to maximize the profits. Just think about the question and write the answer for that. What are the data analytics phases in IOT? And these are the data analytics phases in IOT, descriptive analytics, redictor analytics, prospective analytics. Already we discussed about these different phases of the data analytics in IOT. Let's we will discuss about data analytics, architecture and stack. IOT enables rapid product application development and operation as well as the collection, analysis and sharing of the potentially large amounts of data. Building and supporting the technology stack require substantial investment and a range of new skills including software development, systems engineering, data analytics and online security expertise. Berkeley data analytics stack architecture consists of following components, Apache Spark, Spark Streaming, Shark, BlinkDB and Tycom. Let's we will discuss one by one. A Apache Shark is a high-speed cluster computing system compatible with Adobe that can run 100x faster thanks to its ability to perform computations in memory. Spark provides concise high-level APIs in Scala, Python and Java. Spark Streaming is one of the elements of Berkeley data analytics stack structure which provides a highly scalable, fault-tolerant real-time processing. A Shark, a low-latency query engine that is compatible with Apache but can run more than 100x faster. BlinkDB is an appropriate query engine that allows users to trade off latency versus accuracy. Tycom is one of the components of Berkeley data analytics architecture and it is inculcated in a memory distributed storage system that provides HDFC, HDF as API. And this is a architecture of Berkeley data analytics stack as I already discussed all these things and this is represented in terms of stack. First is Shark, BlinkDB, SQL, Spark, Streaming, GropX, ML, Apache Spark, Tycom, HDFS, HBase, Adobe, Clusters, FC, Apache, MS, OS, ARM, Resource Manager and all components are represented in terms of stack. And these are the references where I prepared PPPT, the first one is a report, second one is textbook that is the Internet of Things and the rest all are.