 Hello everyone, welcome to the session of Monte Carlo simulation and risk analytics. In the previous two sessions, we have discussed the details of Monte Carlo simulation and its application. We have also discussed the basic information of system dynamics. Today, we will discuss the risk analytics part using ADRIX software. We will understand the basic features of this risk analytics software called ADRIX software. It was earlier available by Palisade and now there is another company called Lumivero. They have acquired Palisade and now it is available on the website of Lumivero. 15 days trial version are available, you can download it and you can practice at home. But today I will show you how the software works, the risk analytics software. The fundamental of the software are based on Monte Carlo simulation. We have already understood, we will utilize that particular concept of use of Monte Carlo simulation in the background of the software and then we will illustrate couple of case applications. The case applications could be basic exploratory data analysis and say cost and budget estimation, then NPV calculation as well as you will see different aspects of NPV calculation under risk. And then you know product mix problem which is very popular in manufacturing sector or inventory optimization process and then we will study insurance analytics, decision tree analysis and if time permits we will discuss supply chain disruption and one or two another application say in financial sectors. So, let us begin the software and understand how the risk analytics can be used for prediction or for you know taking a better decision making in the context of business analytics or say business forecasting. Basic Excel you can see here, now in this Excel you can see the address part is not been installed yet. So, what we will do or add in yet, so what we are going to do now we will go install it. So, let us see here you can see that at this 7.5 version is installed in my desktop, so I am going to install it or add in it in my Excel, now if you click the Excel you can see that at this part has come now, but in your laptop or in your Excel it is not available we have to buy it, but you can download it for trial version for 15 days trial version and you can practice. So, let us see how the software works with some basic like you know statistical analysis or exploratory data analysis we will do and then we will go deeper of this particular software. First in the previous session in simulations I have already illustrated a sales promotion problem you can remember that, but let us go detail on that with basic problem of this particular software. Because here I am writing say cost equals to say you know say 200 and say you know say revenue basic problem I am writing now say equals to say 300 say profit equals to revenue minus cost. So, you can see the basic calculation we have done it here, now what I am going to do I will make cost as some risk parameter suppose I am assuming it is a triangular distribution suppose you can define any distribution suppose you know most likely is 200 and minimum is 180 and maximum is say 220 if you wish you can make it you can change this also suppose you want to make it say 150 say right you can do it and say max most likely say 250 you do it 250 say right. So, now you can see you can put any other you know ranges of data as per the actual data set for based on the experts opinion suppose we have made it the cost as triangular distribution we are assuming the investment of the cost or say to some extent random in nature and it is following say triangular distribution for the sake of illustration, but when you have actual data you can feed the distribution and you can also that also you can feed the distribution and you can use that distribution function in your system analysis or the problem of case. Now suppose revenue revenue is suppose following some because you know based on the cost if the cost is uncertain or the investment is varying that means your revenue will also change. So, now suppose we are making suppose your say revenue is following say uniform distribution suppose. So, we have considered it as a uniform distribution here you can see and if you would like to see the you would like to make the changes of that. So, here you can make changes say suppose we will make 250 say we will make little variation of that. So, that max say 350 suppose now look at the input or say investment or cost of your business which is triangular which is following triangular distribution with most likely value of 200 and revenue is of 300 and it has a range of lower side of which is following uniform distribution suppose based on the past data you can see we can put any other distribution also here for the sake of illustration I have put uniform distribution which is following say you know 250 to 350. Now your profit will not be the fixed in the deterministic case it was fixed now it will change because your any every simulation when you run the problem your cost input or investment will not be 200 it may vary with the range of triangular distribution and the revenue will also change based on uniform distribution. So, in every simulations you will put one input of cost and one output of revenue and the corresponding difference will come as a profit. So, that cannot be fixed now. So, we are making as a risk output as per the software whatever the output cell you are mentioning or you are calculating that should be mentioned as a risk output you can put the you can change the name here also by invited comma suppose I am keeping as it is now look at the output cell it is a risk output it is not a 7 minus f6 it is a risk output plus here you can put the name of the you know whatever you can mention suppose you can mention profit say what we will do we will run the simulation suppose let us run it say say for 1000 iterations say now you run the simulation lets you what happens. So, look at we have got the profit of 1000 iteration and here you can see the entire analysis are been or statistical output are been put in the right hand side here. So, in the deterministic case what you have done in the deterministic case you could do maximum the revenue minus cost and the corresponding profit there is no variation, but here what you have done based on the previous data historical data the pattern or the experts opinion or deductive method or you know historical analogy you realize that the cost or say investment follow some triangular distribution for the sake of illustration and revenue which is 300 in general on average, but it also has a variation which is following suppose uniform distribution in that case the route which is profit will not be the fixed it will also vary here we have seen we can see here we have done 1000 simulation we can see all 1000 iteration are been posted here and we can see the mean here the lower bound of your profit could be 12 rupees only and the maximum value could it could go up to 190 where in deterministic case it was 100 only, but the mean since we have done 1000 iteration on an average mean will be whatever the variation of cost or the investment and the revenue are there, but effectively your mean will almost will be almost nearby to the deterministic case the mean the average value expected outcome of your profit would be around 100 here we are getting 100 say if you reduce the number of simulation it might change, but here we have taken 1000 iteration which is quite good so here you can see mean is this and the 90% confidence interval you can see the plus minus changes I will show you later also and look at the mean, median, mode all you are getting through this output only of your profit so this advantage you will be able to get through this very scanalytic software and you can make a better prediction and better statistical decision in making in your organization also and here you can see the standard deviation, but look at the standard deviation is 35 so here the mean is 100, but standard deviation is been coming here because the data the input data cost and revenue are changing so therefore your profit the expected profit now it is not a simple profit it is expected profit the expected profit is coming out to be 100 and standard deviation is coming out to be say 35 percent and you can see the confidence interval here so here 99% 90% 95% all 80% all confidence interval also will project to you given to you now if you like to see the more detail of this particular graph here you can see this 5% means what that there is a chance that your profit may go above 159.4 and how much the chance is 5% and what is the confidence interval 90% 90% confidence intervals we can see 43.8 to 159.4 so in between that profit your confidence interval says that 90% confidence interval says that your profit will lie between 43 in the lower side and maximum it can go up to 159 you are 90% confident on that here 5% outlight outlayer you can see here look at it may go above 159 also it can go below 43 also the chance is 5% suppose you want to see that I want my profit should be minimum 50 say you want to put it here so let's see your profit minimum 50 look at here and the chance is quite high and here you can see it may go down below 50 the chance is 8.1% here so this type of analysis you can do also you can see only 1.4% that profit may go above 170 and if you want to make say you know 100 look at almost 50% look at almost 50% chance that it may go above 49.4% it may go above 100 rupees because we have found the mean at mean as 100 rupees so here if you want to say you know 10% you can see the changes in your confidence interval and the profit range also also one more part let me illustrate it suppose let me keep as it is say say 5% here I am keeping and suppose here I am keeping say 5% say okay now you can see the how much is your risk or value at risk you can also check here also though it is a profit function you can calculate the cost function also if it is a supply chain problem or say manufacturing problem or say marketing sales analysis or say financial data analysis so in that case you can see your you know value at risk in the right hand side and the left hand side so this part you can calculate also how much value of your profit or your cost are above that 5% you can also calculate you can calculate the if you take the average of it you may calculate the conditional value at risk also so many more analysis you can do through this you know at risk analytics software which is available through their website and now if you like to see the tornado graph the among the cost and revenue which making maximum variation to your profit if you click here you can see the tornado graph and here it says that the revenue is making maximum variation to your profit look at here that means that if you keep cost as a static variable that means if you do not change cost keep a fixed value of cost here you can see here look at here the static option I will show you there suppose you know if you click the cost part here look at the risk static that means this part means that look at the range of triangular distribution of cost and look at the risk static it means that that cost will be fixed like deterministic case cost will not be varied similarly revenue can also be done so suppose we are analyzing the profit part and here if you see here I have already run the simulation so you do not need to run the simulation again unless you swipe it and close it now you can see the profit is to some extent is varied or explained by your revenue maximum if you keep fixed cost static or fixed similarly if you keep your revenue fixed in that case cost is making variation to your profit but quite less look at the variation of your profit look at here the variation of your profit through revenue is maximum so you can trigger that so if you can sell the product with little higher margin probably you will be able to generate more profit because it has a maximum variation so you can reduce the downside risk of your profit and you can increase the upside opportunity by increasing the revenue so this type of analysis tournament of graph analysis you can also see you can also you know observe the spider graph here look at the spider graph here so who is making maximum variation to your profit here you can see your actually revenue is making maximum variation to your profit now what I am going to show you so this is the data and we have done the you know basic simulation run of this particular software and we have analyzed a basic concept just one part you see here is a entire you know analysis part you can also actually just if you save it simulation run is already been done you can you know browse your result look at that automatically it is been done and also if you wish you can generate the report in any new workbooks say suppose quick report suppose we will write mention input and output and we will have to see it and we are generating report look at you have generated the report and this report you can actually look at all this part that I have shown you know look at the distribution function the cumulative value and the delta result that you found look at here how many iterations you have done how many input variable you had and then also the number of simulation you have run only one simulation and the number of iteration number of input data number of output data and also you know the duration of your simulation look at the output data sets enter summary statistics you found here with confidence interval lower bound upper bound mean value skewness cartosis look at that here we are getting the skewness cartosis of your profit function also and also you know the range mean median mode all these things you are able to see here and you can also which making maximum variation in your turner graph that also you can see here with the lower right lower side risk and upper side opportunity you can see this output also and you can copy this output and you can paste or save in world or you know ppt or you know in any presentation also so this type of or you know excel also so these are the analysis of the given like aspects the positive aspects of the software that you are getting through this particular analysis of risk analytics using Monte Carlo simulation now what we are going to discuss now go back to the you know main problem we look at these are the few more input and output data that we have analyzed so let me go back to the main file this was the main file here I am going to show you another interesting analysis let me run it again here the profit and if you see the browsing results look at here here what is your mean value the mean value is 100 right after 1000 simulation you can change the simulation number of iteration also suppose if you want to reduce the number of iteration to 500 or see you know 5 to increase to up to 5000 then all this analysis you can do also and you may get more you know mimicking process of the simulation output of your iteration now here if you see say you know standard deviation look at the standard deviation is 35 for this particular data analysis where input data the investment cost and the revenue are following some distribution and profit we have found it here which is also to some extent following little bit of normal distribution but we have found the data here or say in a bell safe car now here you can see the standard deviation remember the skewness cut-off is also you may calculate you can also see here but look at the standard deviation 35 remember the standard deviation is 35 and mean is 100 let me close it now what I am going to show you here is that in this case look at that the lower case of profit is how much 12 and maximum case is how much 190 12 and 190 and mean is sorry standard deviation is 35 remember 12 to 190 let me write down here the minimum value the minimum value is coming out to be 12 the maximum value is coming out to be 190 and mean is how much mean we are not discussing now so mean is how much 100 say standard deviation is how much 35 just note down ok with this data input here what we have done let me use the pen here you can see your cost this cost part look at the cost part this cost is following actually some triangular distribution where you may get a let me click it you will get to know look at the cost so this cost this is the cost function right here you can see it has a you know downside position of so 150 around on an average and upside it can go up to 250 and revenue you can see look at the revenue which is following also from 250 say 350 say now what happens in your 100,000 people understand the essence that I am trying to explain with you in basic simulation what happened in a particular iteration out of 1000 iteration in certain case you might got the lower cost value the lower cost side say say 160 you found 160 so in a particular simulation you have generated 160 through random number look at there I told you through the inverse function scheme of cumulative scheme you can generate from here some random function random number and then from corresponding graph you may get from the x axis you may get the x value the basic Monte Carlo simulation scheme through that suppose you have generated in a particular simulation you have generated 160 and you have generated revenue say maximum how much suppose 340 so in particular case suppose you are generating 160 as your cost and you are generating your revenue say 350 suppose so effectively what will be your profit profit will be the extreme case of revenue minus minimum investment cost so look at 190 profit you are getting right suppose you may get it so these are the extreme cases you are getting in your simulation case so you are making less investment but you are generating higher revenue this is not the true case in practice there should be correlation between investment and revenue that means if you invest less or if your cost expenditure is less generally your revenue should also be less but if you invest more your expenditure is high probably your revenue could also be high so that correlation we are not doing here we are randomly generating cost in every simulation one cost input and one revenue input and we are calculating profit so extreme cases are coming similarly there might be case that you know we are putting huge cost investment higher site of cost but you are getting lower site of profit or revenue higher site of cost and lower site of revenue in that case your profit will be very less so this extreme cases of simulations we are for instances we are trying to avoid how can you avoid that because that is the practical cases so there is a way to avoid that what you do so select them first either you can do so now the cost and revenue has come now suppose I am keeping a 80% correlation suppose 0.8 suppose I am assuming that select a location okay I am doing I am selecting the location suppose here so here what I have done I have created a correlation between cost and revenue how much is the correlation 80% say using the data you can fit it also sample data set suppose here we have generated the 80% correlation between or you have assumed 80% correlation between cost and revenue so higher the cost higher the revenue lower the investment lower the revenue now if you run your profits say after defining this correlation look at what output we are getting look at our mean is almost same look at our mean is almost same but see the output variation the minimum value now it has come to how much 53 and 148 really let me write down here 53 and here we found 148 look at the variation has come down look at and mean also is almost same but standard deviation look at only 17.4 17.4 so look at the analysis of comparative analysis now here come to home and if you put a color you will get to know the differences look at here in this case you have already calculated the output cell where the minimum maximum of your revenue or of your profit with the standard deviation of 35 but once you created a correlation between or not have assumed a correlation between input data and output revenue look at this is sample illustration in practice your problem statement may change your case analysis may change you might have a different case application so in that case this type of correlation you can build among the data sets so which will give you more practical insights about your data analysis so here I have shown you a basic example of cost revenue and profit here so here you can see since you put you have put a correlation between the input data here you can see your variation is coming down and here minimum is 53 rupees and maximum is 148 and your standard deviation look at standard deviation has come down to 17 because the actual correlation between cost and revenue are here this is one such illustration of this particular software I have shown you there are many more such analysis are been or the features are there in the software if you wish you can analyze that through which you can do better risk analytics or better predictive analytics or better business forecasting aspects of data modeling so here I will show you another interesting understanding of excluded data analysis suppose here we have kept a data sets it could be sales it could be profit it could be say revenue it could be say temperature it could be stock price it can be crude oil price it can be gold price it can be your manufacturing process analysis it can be you know in your financial data analysis whatever you can consider right so suppose we have found or some observations of say you know some particular experiment so suppose we have kept a data here of suppose random data I have generated around 50 data how many around 53 data I have generated here say randomly it is observation of any particular variables or say you know of your practical case now if you calculate the average of this data here is the average simple average I have taken and standard deviation I have calculated standard deviation I have calculated now what I am going to show you here is that suppose this data sets this data which is x I have mentioned here is a input say for example say sales right or say you know your temperature which will go to a system input to your analysis of a practical case suppose here system output I have mentioned here as a suppose a function of this input data x say sales or whatever which is nothing but say 2x plus 50 suppose I have written here 2x plus 50 for the sake of saving time I have written suppose the sample value 2x plus 50 which is a system output where this x the column A will go as input now in general what people do people cannot take this data directly right because these are the random data in next time period or next instances or next event what input will go there we do not know in general what people do people take their average and the average value come here as a system input and you find your total our system output as a say 102 simple calculation just mean mean of this data I have kept here this is nothing but simple mean right simple mean say mean of the data I have kept here and I have calculate the system output as it is now what I am going to show you rather than taking the average of the data into system output the system output could be many big problem or complex systems say rather than giving the input data as a mean value I am going to generate a random instances of this data say sales in every practical situation in future or every instances of my simulation. So, I am going to generate any value from this particular data sets in forthcoming instances rather than average value how to generate it first you have to consider that data pattern what type of the pattern the data follow so we have tried to test here look at this data is following some rich normal distribution how come let me run it here. So, data can follow certain distribution that also you can feed here. So, what I am going to do here define distribution say in this particular say okay distribution feed feed the distribution. So, I am going to select the data so I have selected the data and now I will see what type of distribution this data is following look at this data this 53 data I have plotted here and the software suggests that that this data follow normal distribution with the highest priority and then the second priority is the logistic distribution and the weibull and triangular one by one. So, since the software is suggesting normal distribution this data follow normal distribution we will consider this data representative as normal distribution. So, if you write in some cells say let us suppose write suppose here okay. So, we have feed the distribution now this was I have written I have calculated earlier but here I have shown you also live demonstration. So, this is what now the representative of this data set of column A so what we will do we will consider them as a representative of this data set suppose we are going to keep it as some say this one write this cell now this cell I will give as a input to this particular system output or say cost of feed or you know whatever temperature cells whatever. So, this system output will be considering not this average value of this data rather than this dynamic value of the data random value of the data and then we will calculate the system output. Suppose here I have calculated right what I have done now you can see same data to in order to save time we can do it anyway I can show you that calculation here also suppose it is nothing but 2 into x which is x the formula x plus 5 output cell. So, this is now my new x now it is a representative now not the average value plus a 50 right. So, this is different than this this one j say 9 is different than h second. So, this we are going to run now if you run it then you have to calculate you have to consider it as a risk output just click that risk output cell done ok. So, I can also mention it as a system output I can mention here also say system output say suppose I am writing here right system output ok done now let me run it look at here this system output is nothing but a distribution of data of output cell of 1000 simulations say you can reduce the number of simulation also not a matter. Suppose if you want to make a say 500 simulation we just run it again you can generate 500 simulations depending on your requirements or threshold point. So, now suppose you have run this particular data analysis what insights you will get it from here here you could see that your system output is dependent on the column x column a say, but you have not give you have not given the single it put the average value of that data the column number a as a input to the output cell or system output rather or every iteration of your simulation you are actually generating one input of your data set a or data set x and through simulation through Monte Carlo simulation by generating random number and you are from where you are generating from this particular cell right. This cell you are actually coming you are generating into your profit and this cells how to generate random input from there, but from this distribution normal distribution we have already discussed in the basic Monte Carlo simulation session through that you are generating one input every time in this particular cell through Monte Carlo simulation scheme and you are getting this output cell look at here this output cell here now and here you also you can see the confidence interval me in variance standard deviation skewness cut-offs is lower bound upper bound whatever you can make any changes of your data and you can make in a variation of your analysis and you can corresponding analysis we can predict in your management of the case of the presence. So, this is another exploratory data analysis, but there are many more you know I am going to show you one by one I am going to open another interesting file suppose cost estimation. So, this previous files now I am going to close otherwise both software both model will run and we may come up with the previous running file. So, now here at the software is already been installed I do not have to reinstall it again. So, now at the software is here here you can see this is a very interesting case application in specifically in the project management industry cost and budget analysis. Suppose you will have to you know create one of your say project or some you know some construction or whatever or you want to bid something for a particular company in that case suppose you have assumed that suppose this is example of project management industry. So, this way I will elaborate or I will illustrate, but it can be applied in other domain also. Suppose here you can see the land cost the building cost construction of the building cost raw material salaries of the laborers and then it is IT part vehicles part marketing part and the overall overhead cost. So, here you can see the base cost based on the you know your inter budgeting and the cost analysis you may calculate this right, but in India what happened you know this is your total expenditure say you are estimating it before you start the project or go for a bidding or construct the building or whatever the case you are estimating your budget sure you are predicting now. So, now this based on your past experience analogy or historical method or top down bottom up whatever the method you can follow or the you know different Delphi method whatever the method we have discussed. Suppose you have come up with this particular estimation say now what we are going to discuss we are assuming that in generally what happened in India you know once you create a budget and you get the budget sanction, but once you go for implication process or the implementation process you see the cost increases because of the delay because of the labor issue because of strike raw material cost or you know whether it will have a delay in your expenditure generally cost time everywhere in India we have observed that it get delayed. So, that delay part we have captured through the cost estimation here. So, we are assuming that for the sake of illustration that data are following to some extent part distribution in project management it follow part distribution therefore, each data we are keeping as a triangular distribution. So, low downside risk that means downside not optimistic case suppose 2000 is the for the land cost initial calculation is 2000. Now, we are assuming that you can save 10 percent of it suppose in case you can consider no downside like one sided triangular distribution also suppose here we are assuming that that you can save your budget in land or everything. So, 10 percent discount we have kept here most likely as it is and the upper case we are assuming that on an average this is sample elastation only you can change the variation we are assuming that the 20 percent extra cost will be there for land for building raw material etcetera. For every case we have kept 25 percent extra, but you can make a changes of that also. So, if it is a 25 percent extra definitely look at here your land cost will not be 2000 now it will be 2000 almost 500 now because it is been increased by 25 percent. Similarly, downside optimistic case is 1800 because 10 percent less and most likely as it is. So, this is the variation of your data now if you use part distribution you can follow another distribution also for in general you know in project management industry people use part distribution and part distribution formula is nothing, but you know a plus 4b plus you know plus c by 6. So, here we are assuming that each of this you know input data sets are following part distribution and then we have assumed the total output or total expenditure as here look at here this is the sum of them only, but it is a risk output now because all the data we have considered as a you know part distribution how to define the part distribution like you know using the formula a plus 4b plus you know c by 6 that formula you can follow using that you do not have to calculate it here you can actually find this particular formula here the lower point will be this, the middle point will be this one and the here you can see pessimistic part in this particular example, but if it is a profit case it would be optimistic here in a pessimistic point would be this one right. So, if you click it you have created the distribution of that particular data sets of your land which is following part distribution. So, this you can drag now for every one you will be able to create a part distribution and if you take the sum you are actually calculating your entire expenditure right now, but this is a risk output. So, you have to click your add output cell to define your total expenditure whatever here we have mentioned totals. So, you can keep it as it is now these calculations it was available with the data sets and this particular are as it is now suppose here they have put a title inside the output cell as a total project cost. So, we will run this one both you can run parallel. So, now you are going to run say 500 simulation. So, that means, suppose we let us increase the simulation size say up to 1000 iterations we will do it. Now, what happens once we will get the output you will get to know in every iteration of 1000 cases instances you are generating one input of land, one input of building cost, one input of raw material, salary, IT, vehicles, marketing and other overhead cost you are generating one output of total project cost. This is for one instances of different how many variables 1, 2, 3, 4, 5, 6, 7, 8 input data are there for all 8 could you are considering they are following some uncertainty or randomness involved over there in practice you do not know maybe weather maybe labor issue raw material cost delay in the political aspects which anything can come and you might say impact in your total expenditure of your project cost. So, therefore, you are estimating your forecasting your total project cost before you start your project. So, this monitoring process also you can do in between your project progress process like you know you can do the SWAT analysis also and you can also get to know more detail about your you know monitoring part of your budget estimation and the cost. Here we are not going to detail of the project management part, but I am trying to show you that how you can get a prediction well in advance in your hand in your excel or in your pen drive and you can analyze the practical cases through this case illustration of risk analytics. So, now, let us run this case you will get the more interesting insights. So, we are going to run it now. So, you have run see look at there are so many input data we had. So, we are going to we have run that particular output cell total project cost here you can see total project cost and we have found the mean standard emissions q-ness all these things we have found you can create a correlation suppose if you think that if you think that suppose your building cost and say raw material cost are correlated you can create a correlation also no doubt about it I have shown you in the previous example you can create a correlation between building cost and raw material cost. In that case probably your variation will be reduced of your total project cost because you know the randomness will be gone and the correlation between the building and raw material will come that we are not going to discuss now we are just seeing the basic case. So, suppose here we found the total analysis here interestingly here you can see among the output data set which are all following some part distribution and random in nature I will have to see the turnover graph who is making maximum variation to my total project cost look at here building cost building cost is making maximum variation to my total project cost look at here if you keep other variable static look at here total vehicles cost are not that much you know impactful to your total project cost it is has it has a nominal you know impacting your total project cost the variation is very less but building cost is having maximum variation or raw material cost. So, if you can you know reduce this overhead cost the additional contingency cost is too high actually here also you can see for this particular sample data sets that means if you suppose take the building cost. So, if you can trigger the building cost if you can reduce the variation in your building cost or expenditure in your building cost or raw material cost probably you can reduce your entire project cost as much as possible. So, it has a maximum variation you can also see the spider graph here you will see the building cost look at which on the rate one is building cost look at the percentile changes here you can see rate one which how it is making impact your total profit look at here. So, variation is there if your building cost variation increases we can see your total project cost is also increasing with a very high range of lower site and the upper site. So, this type of analysis you will be able to see also interestingly now I am going to show you another interesting part let us close it here. Now, I am going to show you one interesting analysis suppose you here you have almost 7 8 variables right input variables which are random miniature risk is involved in this particular analysis of total project cost. But here also we have through tornado graph and the spider graph we have also observed that building cost is having maximum variation to your you know total output say or total project cost we will have to see the scatter plot between the building cost and say profit. So, what we are going to do we will browse the result and here we will see the scatter plot of your data with say total project cost with your say building cost right just select it and see the scatter plot here you can see the analysis says that look at here if you increase your building cost because of uncertainty here you can see a total project cost is also increasing and entire you know 1000 simulations output are kept here analysis you can copy it and you can paste in your excel ppt and you can analyze this also and you can also see the you know overall bifurfication about your data of your total project with the range with the downside you know quadrant and the upper side quadrant and all this analysis also many more aspects also you can see from this particular analysis of output analysis cell. Couple of I am trying to show you sensitive analysis what if analysis many other thing also you can do using this risk analytics software and further prediction and the recourse action plan also you can analyze actually.