 Hello everyone, welcome to the course of business forecasting. Today we will discuss the basic concepts of data driven decision making or say steps of data driven decision making and essentials of predictive analytics. When you enter into the course of business forecasting, you need some sort of basic understanding of data analysis or the requirement of predictive analytics. So, today we will focus these concepts or the requirements or the say ingredients of data driven decision making, basically the steps of data driven decision making and what are the important aspects that you need to know to start with the predictive analytics modeling or say business forecasting techniques. So, first question would be what is data driven decision making? It is nothing but it is a steps or you can say the methodology where you emphasize the use of data in terms of informed decision making by leveraging the sources of data or extracting the data insights. What you do? You study the data behavior, the pattern and you make a better decision in terms of understanding the consumer behavior or the consumer requirement, satisfaction, competitive advantage, operational efficiency as well as the market trade. So, it effectively through data driven decision making you actually understand the past behavior of the data, the historical behavior of the data and you analyze that through different techniques of data analytics or say you know through a steps different type of you know modeling approach of data driven decision making or data analytics and you make a forecast for the future accuracy of the data, what has happened in the past and through this DDM or you can say the predictive analytics or the business analytics steps, how you can make a forecast for the future. That is all put together I have given a coin called data driven decision making under the context of predictive analytics or you can say under the context of business analytics. Definitely we can talk about business analytics because predictive analytics or business forecasting is one of the pillar of business analytics. So, definitely if you see the right hand side benefits, if you look at the benefits here I have mentioned what are the benefits of data driven decision making or you know essentials of predictive analytics. Effectively you see from opinion and intuition based opinion or decision making to a data driven decision making major aspects here all will be your past data based on the behavior of the data you can make forecast. You might have a question sir what is going to happen that we are totally rely on the past data, but in future if something happens I mean contextual event occurs or event happens then how will data will take care of that. We have a session for that also specific allotted session human judgment the contextual event through experts opinion field experts opinion how the integration of the past data behavior and the future aspect the prediction can be integrated together and you can make a forecast for that also. We will discuss that on that particular session, but today overall majorly in the entire session of business forecasting or the entire module of modules of business forecasting will be concentrated majorly past data behavior their pattern understand the data pattern which will provide the insights to the managers through identifying the trained seasonality and different type of you know random behavior etcetera and make the forecast for the future accuracy or the business models operational efficiency market trend consumer test behavior and the competitive advantage. Look at here drive business decisions with better confidence you will get to know through data will I will show you through different case illustration will get to know how through data analytics also through data decision making you can really take a informed decision and make a bring a better confidence in your decision making or in a business strategy which will not only give you the competitive advantage also it will improve the customer experience starting from the morning when you start taking the tea or breakfast from there itself you know whatever in Unilever you starts with the you know Nestle or you know whatever product Tata tea you starts from that morning company actually starts understanding the consumer behavior, consumer need, consumer requirement for that they need data. Since company has the data or in case they collect the data or you know extract the data or consumer test, consumer behavior requirement etcetera probably if they do the analysis or you know to some extent data analytics techniques and tools they apply on that particular data probably they can get a better customer experience as well as the competitive advantage also. Even imagine whatever I talked about the FMCG products say in Unilever you starts from there in the morning even at the end of the day when you go to the bed for sleep effectively you use the sum of the product of Unilever or any particular FMCG products. So, this is one sector I am talking about that every day we need we use them and their companies are using this data decision making or business analytics aspects to get a better insights and to remain sustainable in the competitive you know market. So, therefore, the first question is that what are the steps involved in data decision making there are three pillar I talked about that first is the descriptive analytics, second is the predictive analytics, third is the prescriptive analytics. If you think about descriptive analytics there our main objective is to understand what has happened in the past it is nothing, but the exploratory data analysis. Today we will discuss all three steps with examples or you know with basic understanding one by one, but first let us think about descriptive analytics. These three put together descriptive predictive and prescriptive we can classify more in a diagnostic analytics also, but overall generally people divide the entire data analytics steps or the analytics like you know business analytics process into three three dimension one is the or you know three pillar you can say descriptive analytics. What it talks about it it says that the in the past how you understand the data that happened or in the past how the data behaved that you try to extract that is called exploratory data analysis throw that or descriptive statistics throw that you can clear dashboard you can use some you know different type of graphs and you can see that the you know basic you know statistical models like you know mean variance data division and you can get to know the pattern or the you know scatter plot etcetera you can get to know the data behavior data pattern that is called simple descriptive analytics. And then you go to the next level once you know the data pattern behavior then you make the forecast the prediction that is called the predictive analytics based on the data behavior past you try to make what is going to happen what could happen. So, that is nothing, but the predictive analytics and there we will use different models techniques of business forecasting. In our particular course of business forecasting we will be majorly focusing on the predictive modeling or predictive analytics. There are another components that is called another module that is called prescriptive analytics. What prescriptive analytics does it actually focus that the strategy of the decision based on the descriptive statistics or descriptive analytics and the predictive analytics predictive modeling or the forecasting plan. Once you have both in your hand then you take a decision that what would what prescription you should provide to the organization or to the company. So, that you can make a better strategy and the competitive advantage that we call it is of what should we do that means what strategy what decision making should come in your hand. There are methods like optimization techniques I will discuss details today. Optimization techniques, simulation modeling, multi-grader decision making and different type of you know game theoretical approach or the you know utility theory or you can say risk analytics models through that you can take really some prescription to the to the data that you have analyzed based on your description of the data the pattern of the data and the prediction model that you have come up through that you can really take a recourse action or the strategy decision making that falls all this process falls under the umbrella of prescriptive analytics. We will focus through these courses definitely we will intricate this you know in our discussions process everywhere we will which we will integrate the concept of descriptive analytics as much as possible as well as the prescriptive analytics also. But our main concerns main main understanding will be predictive analytics or you can say the forecasting in the context of business models, industrial problems or basic case studies related to the management. So, with these three aspects in our mind like you know you can think about you can relate these three aspects with sustainability also. Generally sustainable management sustainability practices are covered under three aspects one is the you know economic sustainability and then environmental sustainability and social sustainability. But these three are linked but not one after another. But here when it comes to the analytics or business analytics or say you know to some extent data analytics three these three aspects of three pillar one is the descriptive analytics, second is the predictive analytics and third is the prescriptive analytics comes one after another or you can say they are very highly correlated one until you complete the understanding the descriptive analytics you cannot start entering into the predictive analytics modeling and then you cannot make the prescription of the data or come up with a strategy. So, all three are interlinked but to some extent it's a sequential once you know better of descriptive analytics you can make informed prediction models or you can come up with a better prediction models with better accuracy. Once you come up with a better prediction you can or you know better forecasting you can think about how much production you should do what could be the you know distribution strategy what could be your you know marketing plan that will come under prescriptive analytics. So, these three are the aspects of data types or data analytics or you can say the business analytics processes. Then this part that I have mentioned in the left hand side effectively can be classified into five six steps of data decision making which we are going to understand through examples of cases say or through different steps of the data decision making process. What is the first step? First step is that understand the case understand the business problem. I will give examples you will get to know and then once you understand the case the business problem that company has given to you or you are going to address that issue but as a part of consulting or as in as a part of you know R&D team or you know data scientist team what you need to do you need to understand the business case effectively. The better you know the business problem or understand go inside of the deep of the problem better you can extend the data and better you can make the model and better you can make the forecast. So, next all steps will be a subsequent steps that the next step is nothing but once you understand the case the business problem the context of the problem next step is that data collection and the data preprocessing. How we have collected the data what are the sources of data? How reliable the data are? How authentic the data are? Whether the ethical practices have been maintained or not? Once you know all these things data collection process I will discuss that you will get to know like different steps of data collection or the process of data collection type of data with the primary secondary. Once you know that and clean the data and make the preprocessing all these things you know remove the out layers these second steps are done now. So, data collection and preprocessing are done now. Now you have to some extent you know your clean data you have you understand the data behavior maybe descriptive analytics you might have used to some extent over here or the statistical you know basic you know data modeling part or you can say the you know exploratory data analysis part you might have done little bit over here. You can use the SQLs and different sources of data I will discuss that. And then once you complete the data collection the preprocessing part clean data with your in your hand or you understand the data you know pattern the behavior of these things then you think about which prediction model which predictive model or the prescription model predictive prescriptive analytics models you need to call that is called the model building. Remember the link the connection that I have talked about the three steps along the three pillar like you know or say types descriptive predictive prescription predictive analytics and how I am linking that with the data different decision making process. Company or the industry people may not be understanding about these three steps or they know these three you know types of data analytics, but they know this you know data different decision making process like you know business understanding the case understanding data collection process and model building they might be using models, but they do not think that this modeling process through the prospects of you know business analytics or say data analytics. So, then model building is the major part which we are going to study as much as possible through our courses of different modules of business you know forecasting. Different prediction models different you know time series analysis, regression analysis, simulation model you know machine learning techniques we will try to cover as much as possible you know Arima models through different type of you know examples case illustration numerical study and different type of you know technical aspects also we will cover under the model building. So, this is nothing, but the design of the problem or the case that you are going to analyze right. And once you get into the model building and once it is done then next step is that evaluate the model testing validation of your model with the data that how much accurate the model is that you have developed or you have prepared and how strategically it is important to the company. So, that called model evaluation. So, maybe you know risk assessment can also be a part of that and to some extent monitoring control can also be part of that and over once all these things are done you testing with the trial data and the training part of done and then you check how much accurate the model is providing and try to implement it. So, this called model evaluation and once that is done next step is the deployment. That means you have developed something you have designed some software. So, you have designed some algorithm. You have designed some model prediction model whatever you know that you have to implement to the company right because you are addressing the case you are doing as a part of consulting. So, therefore, what you have to do that you have to deploy the model to the organizations. So, that is called the implementation pass or you may say that to some extent commissioning part. Once that is done in this case in the general project in the industry what happens once the commissioning are done you might not to go back much to look into that particular project you can see to the another projects. But in data analytics or the data decision making or even say in business analytics what happens you have to monitor the process also over a period of time. That means maintenance part is also very important because every day data are being generated based on the past data or say 3 years data 4 years data whatever the you know model you have developed based on the behavior of the data requirement of the company strategy and the technology available with you you can make you can develop and implement and you can make the commissioning of the project right or the you know model that you have provided to the company. But after 2 years you your data might become absolute past data you might have got new data which can give the bit new trend new pattern new requirement as per the competition is concerned. So, in that case you have to redesign your structure you have to upgrade your model with the new technology new AI process new machine learning techniques you based on the more available data. So, you need to upgrade your model that is called to some extent in this particular steps of DDM or DM or data different decision making it is called model maintenance. Otherwise what happens you know like you can take the examples of Nokia say you know. So, earlier it was very popular, but now they are not been adapted with the you know new technology new demand new customer tests etcetera they become you know absolute in the market now they are not that popular remember you know you can think about say you know Yahoo or you know or could during our times it was so popular as a social media platform. But now people are not using now because of you know new technology with the Google or you know in with say Gmail or say you can say Facebook etcetera this old you know that time it was very popular, but this particular you know couple of examples that I was talking I am giving they are not you know relevant to the current society or the current social media. Now, the new technology new developed new softwares has come which you know customers are experiencing with a better test and better enjoyment. So, therefore, they are available in the market who knows that you know tomorrow some new technology will come and that particular Facebook also might become absolute. So, therefore, what happens in a you know you need to focus about the model maintenance like you know upgradation of the data as the technology you have to put together. So, that you know even if new technology like chargeivity comes also into the market you remain sustainable your business models remain sustainable and you can remain competitive in the market also. So, this all you know model maintenance starting from the business understanding data collection of pre-processing model building model evaluation model deployment and maintenance are the you know six steps of data decision making. Look at the summary of the discussion that I have so far made under the business understanding process understand the market trend and dynamics. Take any case any examples in whichever domain you are working not a matter you think about say insurance industry. Suppose you think about you know say supply chain industry you think about say in a chemical industry anything suppose or the financial industry you think about say insurance industry understand the market trend and dynamics who are the involved over there you know or the competition involved over there or the market trend in the insurance industry or say in a Mitchell fund industry whatever you can you can think about that. Then you know consumer behavior knowledge and the preferences that you have to understand look at as the business understanding increases in your mind also or in your you know experience also effectively you will be able to make better forecast to meet the customer need and the competitive advantage. Then awareness of industry specific factors what are the factors involved particular industry that also you need to know. So, therefore, understanding the business case is very crucial before you make the data decision making without knowing that simple understanding a model and you know implementing it with the data may not be the best suitable decision making or the best strategy of decision making perhaps you need to understand the context in a better manner. Then familiarity with the historical sales data pattern data and the patterns of the data. So, this is called you know data behavior analysis etcetera. And then once you know this all these parts then you understand the the external factors also this is very crucial you can imagine the point that this you know this particular aspects in your decision making. If what external it is like you know today suppose you know for example, say some European countries are fighting each other or say you know Middle East countries are fighting each other that will have a big impact in our crude oil price or you know in our say you know energy resources. So, therefore, if you are working on a project of energy or if you are working in a project of say you know to some extent addressing say mutual fund or say you know insurance industry where in your say NBFC sector you have to think about what RBI is taking decisions. You do not have any control on that what RBI will take the decision what Middle East is fighting each other among each other you do not have any control on that crude oil price may go up go down, but you have to think about that external factor also in taking decision through this particular data modeling process. Then air mass of potential risk which are mentioned already. And then you know last part insights into the competitive landscape and the market positioning where you want to position your product where into position your particular company profile. So, that is also that is the objective right that is the goal you have to be completely in the market and you have to position your product with the best quality as the best requirement and the customer requirement right. So, therefore, these steps are very important as far as business understanding concerned or the data different decision making data different decision making process is concerned. Now, let us come to the essentials of business analytics very crucial part. Remember I have discussed only the first step that means that means, if you come back to this particular you know slides here I have discussed only the business understanding part. So, far we have discussed majorly this part. Now, we will be focusing the name remaining steps one by one, but before we focus on the remaining steps let us understand what are the essentials ingredients or essential techniques or the process that you know as a part of descriptive analytics predictive analytics and the prescriptive analytics. Let us go to that particular slides you will get to know. Look at here I have now categorized three definition three term descriptive analytics predictive and the prescriptive. Look at the data types data types data collection process or the data that you have in your hand whether it is a primary data or there is a secondary data that you need to understand I will discuss that in the forthcoming slide. Then once you know the data pattern and then you need this kind of couple of you know techniques or you can say the tools. If you know all these tools and techniques I have put couple of there are many more which I will discuss in the forthcoming sessions even today also I will show you may be in the after the break I will show you. You will get to know that the more you know the techniques and the tools of data analytics or the business analytics probably the more you will be the stronger. So, therefore, if somebody says that you know I know the business analytics it means that or the techniques of data decision making or you know data science you have to be kept in mind that whether you are aware of this techniques or not. If you know these techniques effectively with analytical concept technical aspects and with the data modeling and the prediction process also or the strategy decision making also like inference drawing process also then only you will be able to become a stronger data analyst or the business analyst. So, therefore, these are not only a limited topic of data science or you know say business analytics there are many more, but here I have listed couple of important techniques that you need to know that is called data collection and preprocessing of the steps of that and the exploratory data analysis which may fall under business you know descriptive analytics or the statistical you know data modeling part and then you know to some extent quality and decision making under uncertainty also which can be a part of all three then statistical inferences which also can be a part of all three then supervised and unsupervised learning. I will discuss in the session of you know introduction to machine learning you will get to know even today also I will give basic information about these two aspects also then classification and clustering very important you know dimension reduction different type of K-means clustering you need to know them the more you know the this parts you know supervised and unsupervised learning and the classification and clustering process the more will be stronger on you know data science part or business analytics expert. Then regression analysis as a part of predictive modeling which we will be discussing through our you know all these courses or the modules and the various types of time series modeling whether it is you know moving by moving average methods whether it is a you know different exponential smoothing models seniority we will discuss about ARIMA models different type of all type of time series forecasting we will be covering through this particular course of business forecasting or essentials of predictive modeling then different type of machine learning techniques also which machine learning is all put together can be put under the umbrella of machine learning but I have given a different term that machine learning and the AI process the most important trend today in the data science domain and then text analysis when you have a say twitter data or different type of you know qualitative data or the experts conversation etc. or people are writing in the different text so that how can you extract that and you can make informed decision by you know doing some sentiment analysis that can be important ingredients for data science or you know part of data analytics and then you know decision analysis is very crucial part where you know all these last four parts are majorly you know including the statistical inference are nothing but the prescriptive analytics these falls effectively you know this falls here like these last four five points including statistical inference like of decision analysis say Bayesian analysis or say decision tree random all these things comes under you know prescriptive analytics optimization techniques linear optimization non-linear optimization mixed integer programming dynamic programming which is look at dynamic I talked about I am talking about dynamic programming it looks very you know basic concept of the mathematical optimization models but it is one of the important techniques of you know AI process some different search algorithms say particle sum optimization or say you know to some extent say and calling operation genetic algorithm which are all important ingredients for you know recommendation system of AI process because that they are all recommendations are based on the search algorithm but customer are searching and how this searching process or the customer behavior you can capture through the algorithms and then accordingly accordingly you make the forecast or you recommend the corresponding product or the corresponding the movie or corresponding the ad it is nothing but are nothing but the search based on the search algorithm so that also comes on the different optimization and the decision making process of decision analysis process different multi-grader process also for the prescriptive analytics also like you need to make ranking of different supplier ranking of different project sites you can use the multi-grader decision making based on the data group decision making you can take a decision which is all which are all strategy decision making also you can use multicolor simulation continuous version of simulation discrete version of simulation I can you can actually address different complex problem through a simulation process and you can take a informed decision or you can make a prediction or prescription through simulation process also so these are the major major you know modeling or the major you know chapter that you should know as a part of decision making or the business analytics or say data science if you know all this if you are expert on all this topics that I have mentioned here again I repeat these are not the limited there are many more techniques which you may need to understand on read or you know you need to gain knowledge about that on the whole these are the major important topics that you should know as a part of business analytics or the data decision making or say you know data science. Now let us enter into the essentials of predictive analytics so far I have given a holistic approach of data decision making and also you know essentials of you can say overall business analytics like I have talked about descriptive analytics predictive analytics as well as the prescript analytics and different techniques also I have discussed. Now let us focus about the core aspects of our course that is called business forecasting we will focus now to some extent or you know in depth of essentials of predictive analytics but if you talk about what is predictive analytics what do you need to do actually it is in the it is an advanced branch of analytics that uses data statistical algorithms techniques machine learning techniques which I have discussed I will be discussing further again to identify the future outcome help make the predictions in making predict better predictions and to make a decision for the future and the you know understand the consumer behavior and and plan a strategy accordingly. So, therefore predictive analytics are nothing but a decision making through data through which you understand the past behavior of the data and analyze the pattern of the data and make a forecast for the future of that particular event that you need to take a decision or need to take a strategy and meet the customer's expectation in a better manner. Now couple of you know you can see couple of dimension or you can say couple of applied area that I have mentioned here where you can really use predictive analytics. First look at the informed decision making here it provides the organizations with insights that enables informed decision and the data decision making by which you can help in staying competitive as well as adapt to change conditions. Although overall you know informed decision making it helps to take an informed decision making then risk mitigation through that you can really throw data effectively predictive analytics effectively you can meet the risk assessment in a better manner also you can you know mitigate the risk and challenges also whether it is a finance whether it is a marketing whether it is a healthcare whether it is a you know project management whether it is a even if you imagine project management industry. In project management industry risk mitigation is one of the most important part and in that case if you do not mitigate risk with appropriate data analytics techniques or the predictive analytics models you are forecast about the budgeting you are forecast about the cost you are forecast about the timing will be wrong effectively what will happen your model will not be resilient. So, if you want to come up with a better resilience models with better recourse plan or a better you know project execution strategy with appropriate with to meet the timeline you need to mitigate the risk and the associated challenges effectively. For that you need to come up really some good prediction models for example, say you know cost estimation process you can think about earn value management you can think about you know some efficient frontier techniques you can think about you know decision tree analysis which will help you in mitigating risk. Then resource optimization it is also very important because you know you have to optimize your resources effectively and the allocate the resource in effective manner right. So, it helps actually optimize resource allocation by providing accurate forecast of the demand, enabling business to manage inventory production and workforce effectively. So, the effect effectively when it comes to the optimization process that is prescriptive modeling, but you need to think about effective forecast right. If you do a effective forecast of your demand like you know to some extent inventory management planning the safety stock planning probably you will be able to do better about your you know procurement planning you can do better your production planning and the managing of the workforce also. So, it is effectively you are doing optimization or resource optimization, but based on your prediction right based on your demand planning etcetera. So, this is also very important part of predictive analytics. Then you know operational optimization also it is effectively the better if you know if you go to any supply chain courses or you know in your operational course strategy courses you will get to know that effective demand planning or effective forecasting whether it is a manufacturing sector or supply chain planning or you know you know to some extent optimization of operational or the strategy of the whether you know tactical process, strategical process and the operational process you need to focus about the you know better planning through prediction models. In any segment of you know supply chain you think about whether HLN1 or upstream supply chain or downstream supply chain you need to focus about your better prediction models. So, that you can optimize your operation and the better coordination about and the bringing efficiency into the model. Then customer retention unless you do a better prediction model you cannot maintain your customer retention. So, chain management is one of the most important aspects in marketing or you know even in business strategy domain. So, therefore, it helps in identifying factor that contribute to the customer chain by understanding customer behavior businesses that can improve the strategies to retain the valuable customers. So, predictive analytics or the data analytics perhaps in most important most efficient most required for you know customer retention. If you do not use data analytics, if you do not use predictive modeling perhaps you will lose the market share. So, most of the companies whether it is in Amazon or whether it is you know our Indian companies also you know everybody is utilizing this data analytics in their decision making. So, that even what customers are going to give the order in the next week you know Amazon is doing some AI process in the data analytics techniques. So, that they can understand that this type of products customer is going to give repeat it order in the forthcoming week. So, let us prepare or store this particular products for this particular customer. That means, they know the data behavior of the customers behavior of the and their past you know buying pattern based on that using data analytics they are trying to make forecast that this customer is going to give this particular order in future week or forthcoming week. So, let us plan or keep the required documents or products for him. So, that you know immediately we can deliver the product to the requirement of the customers. So, this is this way you can improve your customer retention also. So, this is one of the important aspects of predictive analytics which will help you in maintaining your customer retention. There are many more applications like I talked about insurance industry right or the banking sector everybody is utilizing their you know data analytics techniques over there. The those who are utilizing data analytics they are actually you know efficient efficient or maintaining their market share otherwise they might lose the market share also. So, let us take a break now and after the break we will continue the topic of essentials of predictive analytics and the steps of data