 Hello everyone, welcome to the course of business forecasting. Today as a part of introduction to business forecasting, we will discuss the importance of business forecasting in management and different application areas where business forecasting people use and the detailed course outline. So, so the first question would be what is business forecasting? Business forecasting is a technique through which you can analyze the business scenarios, the past data and you can get insights of the data and you can make a better decision making or informed decision about the organization's prediction and future action plans. Using business forecasting techniques and the algorithms, you can actually understand the interrelationship between the different departments. For example, suppose how the sales are being estimated that will have an impact in your marketing plan. How the demands are being estimated that will have an impact in your inventory decision making and production planning. So, it is all about the prediction using the past data and the expert's opinion and implicating of that concepts through different techniques of business forecasting, you can actually remain sustainable in the competitive advantages. Now, you know why we need to study the business forecasting. So, here I have created three different segment you can understand that you know what are the different domain or different aspect that you can actually address through this course of business forecasting. In the recent development of predictive analytics and machine learning, you can use the techniques of data science or say business forecasting which will study throughout our courses and you can analyze the historical data and you can make a better decision. So, this is what the start of business forecasting then what insights you will get. You can manage the risk, you can manage the resources effectively whether it is a manufacturing system, production system, sales, marketing everywhere you need business forecasting. In the project management domain also it is one of the important aspects of planning of your time, cost, budget, everything. Gantt chart preparation through which you can make a better planning of your day-to-day activities. So, they are also unit forecasting. So, in every segment resource allocation say supply chain cost optimization, every unit demand planning, for marketing unit sales planning. So, everywhere you need some sort of business forecasting aspects then only you will be able to take a better insights or better decision about the organization. Look at the last part using the data of organizations, past data organization and the techniques of forecasting or methods of forecasting. You can actually analyze the data behavior as well as the scenario analysis you can do through which you will be remain sustainable in the competitive advantages. What opponents are doing, what competitors are doing that also you can analyze by analyzing the data and the market trend through the process of business forecasting and you can help in designing a better decision making planning for the organizations with the help of different techniques of business forecasting. Now the application area, to some extent you know here I have listed couple of area whether it is a retail, whether it is a finance, manufacturing, healthcare, energy, telecom you know supply chain, everywhere transportation system, media everywhere you need forecasting. So, in fact whatever the business domain you go or industry engineering aspects you capture or you analyze you need the forecasting first. This is the foremost important aspects of business planning and the decision making strategies. Think about retail, if you cannot calculate the demand planning or say you know the sales planning effectively how will be able to take a production or you know storing of inventory or warehouse management you cannot do it right similarly in finance. If you do not do proper you know budgeting as well as the stock price analysis or say you know prediction about your future growth or net present value etcetera you would not be able to take a better financial planning decision also. Think about healthcare, perhaps in India we have not done much research in healthcare as well as the analytics in healthcare aspects. If people do more research on analytics in healthcare using different techniques of predictive analytics or say you know business analytics probably this sector will grow faster and they will remain more practical or you know consumer oriented by analyzing the data analytics in their day to day business process. Here I have listed three segment like demand forecasting. This is common it can be whether it is a banking sector, whether it is a finance sector or say supply chain sector or even in marketing. These are the common aspects that you need to capture through business forecasting courses or techniques of business forecasting. What is the first aspect that is called demand forecasting. If you can do a better demand planning or demand forecasting think about say ACN pens. ACN pens use data analytics at most Hindustan Unilever. In our day to day operations we use products of Hindustan Unilever right. So, the consumer test can be assessed through data how consumers are consuming the products. If Hindustan Unilever or say ACN pens can use data analytics in their decision making process probably they are using it maybe because of that they are remaining sustainable or top in the competitive advantages. Similarly, sales forecasting. In marketing you know if you do not do sales planning effectively you cannot do the marketing plan also as well as the strategies of marketing. So, sales forecasting, demand planning even earning. earning is also very important right. So, how you will make your enter earning and the future growth of the sales forecasting. So, for that you need the past data and you have to analyze the data pattern as well as experts opinion. If you can integrate them then you will be able to do a better sales forecasting or sales planning of your organization. Another important aspect is called the financial forecasting. So, budget calculation, your forecasting of say you know different costing part as well as balance sheet preparation, net present value calculation. In one session I will show you how the net present value can be forecasted using the different techniques of business forecasting under risk environment. We will discuss that. So, these are the couple of applications area in different domain sectors through demand forecasting, sales forecasting and financial forecasting. In forecasting I have raised that questions why you need to study business forecasting right. It is all about prediction and analyzing the data pattern their behavior and making informed decisions. So, keeping that definition in mind what are the approaches that you need to follow in any business domain, whether it is a marketing, whether it is a sales, whether it is a project management, whether it is a you know construction domain wherever you go you need forecasting planning right. For that you need some approaches to adopt the forecasting to calculate the demand to calculate the cost to calculate the sales etcetera. So, company follow couple of approaches. Here I have listed couple of popular approaches through which company assess their sales demand and the financial planning forecasting planning. What are the approaches? First is the bottom up method. I can give one example for each of them you will be able to relate better about these techniques or say you know approaches. First one is the bottom up approach. It is think about say project management industry. In that case what happens you know the bottom level of the organizations like say you know the project manager of the unit of different divisions they estimate the cost, they estimate the budget, they estimate the forecasting related informations like demand planning, sales planning, the representative. They calculate all these budget and then that means segregated. From each unit you can segregate the data, the forecast of the you know total budgeting plan and then you can after you know integrating all them after adding all them you can pass it to the top management. Then top management will take a decision after having a group decision making. So, this approach is called bottom up approach. From the bottom level of your organization in a hierarchy you goes up to the top and the top management take the decision. But in top down method it is other way. Here top management take the decision first. How much cost should be allotted for a project? What could be the tentative timeline for completion of the project or say you know what could be tentative demand plan, sales plan whether you should go for a new product launch or not. Top management will take a forecast and take a decision about that. And then that budget, that time, that cost or that forecasting planning will be passed or the amount will be passed to the bottom level through different divisions and through your different channel. Think about historical analogy. This is also very important aspects historical analogy and many organizations follow that actually. So, here what you depend on the past data and whatever has happened in the past that will be a reference for you. And accordingly you can implement with the new project or the new timeline or you know new product line extension you can follow similar approaches. So, it is called historical analogy. Now deductive method is other is different than historical approach. In deductive method you actually do not rely on the past data. Here experts says that or experts believes that whatever has happened say in Nagpur whatever the project you have executed that may not be relevant in Bombay. Because in Nagpur the rainfall part is less. So, if you the diameter of the pipe size of pipe may not be the same in Bombay. So, here you need to you know increase the diameter of the pipe for water flow in construction of a bridge say. Because here you could see that there will be heavy rain during Mansoons. So, you need to change the policy. So, here you know to some extent experts says that we will not rely on data rather we will rely on based on our experience. And the current situation current situation what they demand accordingly we will modify and plan our business planning and the strategy and the forecasting plan. So, these predictions under deductive method is also very popular. But here everything depend on the experts experience and there to some extent confidence level and different principles they have learnt throughout their career and that they implement through a deductive method. So, these are the four approaches that I have mentioned here. But remember these are not the techniques. These are only approaches like bottom up or say top down or historical analogy or say deductive method. These are the execution process, the decision making process by the organizations, but not the technique. In order to execute these four, five approaches whatever the approach you follow you need some methods, you need some techniques that you need to implement through this or the managers need to adapt these techniques through which they can execute these processes. So, what are the techniques as a part of business forecasting you need to learn or as a part of predictive analytics you need to learn that we are going to discuss throughout this course. So, what are the techniques? Here in this slide I have put almost all the topics of business forecasting that we are going to cover different techniques, different methods, different approaches that I have mentioned here. First I have categorized the entire course into two components. One is the subjective approach another is the objective approach. In subjective approach we will discuss the qualitative aspects of forecasting like group decision making, Delphi method etcetera. In quantitative approach or objective approach we will study time series methods almost all popular methods that is available in industry and the you know in the science of literatures. We will cover all of them in time series module and then we will study the regression analysis techniques also detail of them and then we will study the simulation approach of forecasting also. So, first we will study the essentials of predictive analytics or say essentials of business forecasting and then different steps of data different decision making. We have allocated two sessions on that from the next class onwards we will concentrate on that understanding and then we will discuss the introduction of machine learning as a part of predictive analytics where we will cover two aspects the unsupervised learning and the supervised learning. As a part of business analytics or say business forecasting you should know the minimum information of machine learning that we will cover in one session detail about it after giving the information about machine learning in the session of data different decision making. And once that introductory information of essentials of predictive analytics, steps of data different decision making and what are the techniques that you should know one should learn to become expert in data science or say you know business analytics domain. Once that part is over then we will enter into different techniques of business forecasting. The common techniques the popular techniques of business forecasting one should know to become expert in the industry or to apply these techniques in the industry and solve different case applications or practical problem. So, what are the techniques as I mentioned under qualitative methods we will cover four methods one is that you know the market survey method and then jury of experts opinion or say to some extent you know panel of experts and then sales force composite. So, here you can see the market survey what happens you know you create Google say it or you know directly go to the consumers and interview them and collect the data from them. So, sometimes geographical or the demographic issue comes or the sample size create a problem in terms of your decision making final decision making and the predictions. Because you know age group could be a issue, gender could be a issue to some extent you can see the location the demographic aspects could be a issue in which month you have collected the data that could be a issue and then only you can make the forecast. So, consumers of market survey is a very popular method it is evergreen method, but sometimes it has a limitation. But when you go to the sales force composite what happens here you know you actually rely on the representative who actually manage or handle the consumers or say retailers. So, if you can take their opinion on a holistic manner you are able to understand the sentiment of the people. So, through them you can capture the demand or the sales and accordingly you can make a predictions. So, this is called sales force composite method we will discuss detail of it then jury of experts opinion the panel the committee which is very popular in any segment you go whether it is a business forecasting business aspects management or the politics or say you know even academic domain. Wherever you go the organizations form a committee and the committee sits for a brainstorming session and they come up with the decision. Under that they might use some different approach like bottom up approach I talked about top down method historical analogy different deductive method you know say time series techniques causal models like regression analysis etcetera they can use, but it is a group decision making right. And then the last option is the Delphi method which is very popular method the superset of enter all the methods of you know qualitative approach where you know different layer of iteration comes into effect. And different groups are being formed and each group come up with their forecast in the initial level then there will be a panel coordinator the panel coordinator will give feedback to each group why you are deviated from the other groups. And then each group will come up with their revised forecast by using any techniques of time series or say you know forecasting approaches or a machine learning approaches. Then after a couple of round of iteration there will be a consensus among the groups. So, when you have when you take a decision throw iterative process among the groups where the identity of the groups are not been disclosed and if you come up with a consensus this process are called Delphi method and it is very popular in industry. We will discuss detail of Delphi method in one session as a part of qualitative approach of forecasting methods. Then we will enter into the time series method. In time series method there are dozens of forecasting techniques are available but what are the most popular that is been used in industry that we have listed based on my experience and the different application domain I have listed here couple of them we will study one by one in our forthcoming sessions like name method moving average methods under moving average methods there are different like you know techniques also like you know simple moving average then weighted moving average which is better than simple moving average and then there will be another extension called the exponential moving average. In industry specifically in the stock market people use every day the exponential moving average they do not use simple moving average. So, how the exponential moving average works that also we will study with applications and then we will enter into the exponential smoothing method. In exponential smoothing method you know there are different approaches like single or simple exponential smoothing approach then double exponential smoothing then triple exponential smoothing when you have basic data you can use simple exponential smoothing and then when you have a trend in the data you can use double exponential smoothing in that case you can use hold model right to analyze that trend and then if you have a trend and seasonality together then you can use winter method or say decomposition method. So, both method will discuss winter method and decomposition method when there will be seasonality analysis. In seasonality analysis you need to calculate the index of each season and then you have to aggregate it and then you have to do the decentralization process and then you have to make a forecast for the forthcoming years or forthcoming quarters. In India you know most of the products earning sales etcetera follow seasonality because in India there are various seasons like you know say quarter to quarter basis season you can define like quarter one quarter two quarter three quarter four you can define season basis like like Mansun season summer season winter season this way also you can you know festival season you can define different type of seasonal aspects of data. In India it actually happens most of the company follow seasonal pattern and their sales and earning comes under seasonal basis. So, how to analyze all this type of data with different techniques of seasonality analysis will understand that like quarterly average method winter method and then decomposition method will study all of them and then we will discuss the detail of ARIMA model which is one of the most popular technique of machine learning techniques or say predictive analytics will study detail of it under ARIMA method like auto regressive integrated moving average process what are the steps of air process what are the steps of MA process what I stands for the differencing process we will understand detail about that and then we will understand why SCF and PSC function are so important in in analyzing ARIMA models. So, we will understand the technical aspects as well as the application aspects of SCF function and PSC function through correlogram diagram which will be one of the important assets for you in managing different time series models of say in a practical examples through ARIMA process. Then one of the important aspects that I I liked most that is called human judgment in time series. In time series what happens you take the past data the historical data you analyze their pattern behavior by different techniques here you can see the graph here you have a you know seasonality as well as the uptrend also. So, which model to select for this type of data pattern to make a future forecast you have to select the model right. So, these are all time series data whether it is ARIMA or decomposition method or hold model winter method moving average method whatever method you follow it is actually the based on the past data you understand the data behavior and you make the forecast for the future. But what is going to happen in future that your time series data cannot capture only field experts can understand that. So, if you can integrate the future event through a contextual information and if you can integrate that with your time series forecast and if you can revise your time series forecast that forecast could be very effective. For example, we do outlayer right we remove the outlayer in time series data. So, why we remove the outlayer because we feel that in the past particular years or particular month something wrong happened and because of that there is a spike in sales or say demand or say down in say because of corona or whatever or say Ukraine war or say you know say Palestine war maybe because of that there will be a spike in say stock price or energy crisis. So, next year it might not be there. So, it is a outlayer we delete that if you can do that for past data. So, why can to integrate the future event also if you know that something is going to happen. If you know that you know for world come match final match there will be a huge crowd in Amidabad city or say if you know that you know because of Ukraine war there will be a spike in crude oil price. Your past data does not know that and you cannot make forecast based on that and experts only know that. If you can integrate that concepts along with the time series data and we can modify the revise your time series model and the revise forecast could be much more effective than the basic time series time series data dependent forecasting process. So, that we will study in one session under the heading of human judgment or experts judgment in time series forecasting. Then we will discuss different regression models we call it is a causal models also econometric models will study detail of simple linear regressions and its ingredients like you know measure of accuracy, coefficient of determination, R square, standard error, confidence interval, prediction interval all these things we will study detail and then we will enter into multiple regressions. In multiple regression we will study the basic multiple regression, we will study the multicollinearity, we will also study the normalization process under multiple regression. So, all three aspects of multiple regression we will study and then we will enter into the logistic regression. In one session I will discuss details about logistic regression in simple linear regression or say multiple linear regression what we do? We actually consider the independent variable and dependent variable and we find their causal relationship right. How much the dependent variable is explained by your independent variable? There could be one independent variable, there could be many independent variable and that process are called the basic regression. But in logistic regression your output variable will be dichotomous kind of thing say yes or no not the actual value of y will come or the output dependent variable will come or you need to forecast. You need to forecast whether there will be yes or no whether there will be say 0 or 1 kind of situation. So, in that case example could be say you know in banking sector whether the company would like to or bank would like to give a loan to a you know customer or not or say insurance sector say you know whether the insurance claim is a fraud claim or right claim. In Titanic example you know whether some person whether you know you can use the machine learning technique or say as a part of logistic regression and you can predict how many people died or how many people have survived. So, this type of you know different examples or popular examples are there which we will study couple of these examples through the session of logistic regression. And then we will study the detail of Monte color simulation and business forecasting. Under Monte color simulations we will understand the steps of simulation process and then the application of that both in discrete case and continuous case through system dynamics approach also as well as the you know Excel illustration. In discrete case we will study different applications in say voting line analysis. In continuous case we will see different demand planning or the sales planning or say stock price or sales price fluctuation we will study all of them through dynamic nature of business environment under the context of Monte color simulation and system dynamics approaches. Then we will enter in the last session or say last module we will enter the detail of risk analytics part using simulation as a part of business forecasting. In that module I will give the introduction of the software called Adrix software or say Palisade Decision Tools software. Under that different statistical analysis can be done whether it is a exploratory data analysis whether it is a predictions or whether it is a prescriptive modeling also. So, all three aspects can be covered using this software we will study detail of it and then we will analyze couple of case applications whether it is a cost estimation whether it is a product line extension in manufacturing sector whether it is a NPV calculation under risk or say insurance industry applications like insurance analytics couple of applications we will study through the software as a part of risk analytics using simulations and essentials of business forecasting. Now, let us see what are the steps involved in making a forecast. First you need to understand the context of the problem and then once you know the context the data collection process and you know the time horizon that you need to make the forecast whether you look to forecast for say short term prediction or say for some long term prediction or the entire annual year. So, that you need to find finalize first like you know horizon of the data and before that you know the context of the problem the objective that you need to set first in which context you are making forecast whether it is a you know financial problem or stock price prediction or whether it is a weather prediction or whether it is a gold price prediction or say you know GDP calculation. So, context you need to understand once you understand the context then you have to find the time horizon. Once you know all these parts then data collection and the preprocessing is very important like different components of time series you need to understand and the causal relationship between regression like you know independent variable and dependent variable that you need to understand effectively. So, once these first three steps are ready then you can think about selection of a model. So, all these models we have discussed right based on experience or the practical illustrations you will be able to understand for a practical problem which model to select. So, once you select the model that in a separate session we will discuss about it. So, how to select a model for a particular context and which model to select. So, that aspects need to be covered in a effective manner so that you can find a bridge between your you know model as well as the practical case application that part we will discuss in a separate session. And then once you select the model model selection are done then model evaluation model training and testing you have to do and then once the training and the evaluation part are done you found that there is a less like you know error and better measure of accuracy or higher R square or lower RMSE once you finalize that with your data and the model and the basic training and testing part are done validation are done then next is that implementation we call it is a forecast or to some extent you know commissioning part. So, that part you have to do. So, once the forecast are done the last part is also very important the monitoring and controlling part because whatever the forecast you do based on the current data right once you have a more data new technology comes into effect or new machine learning techniques are there with your hand probably you need to modify your forecast for better accuracy. Now the question is that what are the best practices that you should follow? This is one of the unique aspects that you need to understand that what are the best practices available to remain sustainable or remain competitive in the market. You can follow any business forecasting technique right or you can understand the detail of business forecasting. But you need to understand how to remain in the top you may be a good doctor you have it so much of knowledge but you do not know which medicine to give for which patient. So, that the understanding is very important and how will enhance that knowledge or experience through best practices right what are the best practices steps or some sort of factors here we can see data quality and the accuracy this is very crucial if you maintain the data quality and the accuracy the completeness of the data the auditing part the cleaning process of the data regular auditing very important if you can do that probably you will be able to make better forecast or you know to some extent you can do a better scenario planning here I mentioned scenario planning you do not know what is going to happen. So, that situation you can manage in a better manner through scenario planning with better accuracy if you maintain the data quality effectively. Then advanced technology and the analytical approaches in these days what happens you know every day we get to know that different type of new techniques of machine learning AI process are being coming up right or data analytics techniques. So, if you can understand these techniques and if you can integrate with your data or you know if you can upgrade your software probably you will remain sustainable in competitive advantages. So, these are the best practices you have to follow actually here you can see the consider the external factors. Tomorrow if RBI come up with a new policy of report that will have an impact in your NBFC forecasting planning right growth or say you know metal sector or say you know banking sector. So, how to analyze that even IT sector also if US take a decision US take a decision that will have an impact to the IT sector also. So, external you have to be very careful about the external factors also. So, that you have to keep in mind what is going to happen tomorrow in Middle East you do not know. So, that will have a big impact in your say crude oil or say you know chemical products. So, you have to very careful about the market condition you do not have any control over that, but you can analyze that through experts opinion and you can modify your business planning effectively. So, these are called best practices you have to you know look at the economic conditions, trade market trade and all these things external factors you do not have any control, but you have to be careful about your business planning as well as you know modification of your forecasting approaches. Then collaborative forecasting whether it is internal division whether it is a marketing team, sales team, project management team you need to coordinate with each other. Then regularly update your models that I have mentioned that you know since new data are been coming or new technology have been developed every day. So, you need to make sure that you should make happy of your stakeholders, consumers etcetera by updating your models otherwise model may become obsolete. Then last part monitor and evaluate the accuracy this is very important what are the KPIs involved in your data right or in your model forecast that you should understand if some parameter are making very impactful outcome I will show you in the risk analytics session you will get to know maybe particular one parameter on one variable will have a maximum variation to your total cost or total profit. So, how to capture that particular how to identify that particular factor or parameter k key performance indicator which will have a big impact to your overall performance of your system output system. So, you need to capture that you have to monitor that. So, these are the couple of best practices. Now, when you go to different sessions as I mentioned we will cover many models business forecasting models whether it is a name, method, time series, simple times, moving average process, weighted moving average process, exponentials, moving average process, exponentials, smoothing, hold model, winter hold model, decomposition method, regression, multiple regression, you know Arima model, simulation different models will be covering in different sessions with practical applications and also illustration of them through Excel. So, here I have shown you one screenshot of one session, one such session how we will illustrate different models. We will select different data, this data you can see, this data follows seasonality and trend because the seasonality and trend. So, there is a trend as well as the seasonality also. How to select the best model for them and how to make the best prediction which will follow seasonality as well as the trend. So, your forecast of a data would not be like this or would not be like this. This forecasting will be like also in a similar pattern. For each and every sessions we will take some practical problem or say illustrative examples and we will study them using Excel. So, that will be an integral part of each and every session of our course. When we will open the software, a badric software, different analysis of say you know statistical inferences you will draw also like you know turner to grab analysis, scatter plot, how to feed a distribution of the data or what type of distribution the particular data is following the trend that also you can analyze, what could be the confidence interval, the standard deviation, the skewness, cut-offs is all you know explorative analysis, predictive modeling as well as the prescriptive analytics also will cover during the module of risk analytics part. So, these are the overall you know aspects of course content and the course planning that I have discussed today and in the forthcoming week onwards one by one we will execute all these techniques forecast methods of you know business forecasting in our course. So, these are the books that you can refer the text book that we are going to recommend is the business forecasting by John Ankey and Dean Wieser. The remaining two books are also very popular like you know essentials of business analytics as well as the business analytics by you know Albright. This is very good book as well as the essential of business analytics are one of the very popular book like you know descriptive predictive and prescriptive all part are been covered by them, but we will be covering majorly the predictive analytics part as a part of business analytics. So, this is the overall introduction, the importance of business forecasting course, the course outline and the implementation of different modules of business forecasting techniques using Excel and the reference text book. So, this way we will cover the entire course of business forecasting and today we have given the introduction about the course and the course outline from the next class onwards will enter into different topic or different module of business forecasting. Thank you.