 or we can recalculate our own threshold values based on the values for the last year or last 11 months. So for the session today, we'll take the same example as yesterday for DPT doses administered to children and comparing it to the DPT doses threshold value. And this threshold value will be based on the 11 months prior to the last month. So if we have to calculate the threshold value for April 2021, we'll consider 11 months receding from February 2021. We will skip the value for March and we take 11 months February backwards. Now to do this calculation, we need a predictor. So as we discussed yesterday that predictor configuration and calculations is little tricky and it needs some practice. And our focus on this session is more towards the validation rules and not the predictors. So I will just show you the predictor that we have already created for this threshold value. I also run through quickly run through the documentation for predictor which you could study later after the session. And if you have some specific questions, we could answer them on Slack. But for this session, we keep our focus only on the validation rules. Now where do we create a predictor? So in the maintenance app, we have something called predictor. I click on this and I see the list of already existing predictors. For my validation rule, I will be using the predictor for DPT, hepatitis, doses, threshold value. Let's look at this predictor now. So it has a name, we can add short names, code description as required. And then there is something called output data element. So we have already created one data element for storing the value from this predictor. So if you look at this, this predictor gives you the list of data elements and you can select the data element where you want to populate the value of the predictor. Now for our validation rules, we will be using this data element. So basically a predictor predicts the value and stores it in one data element, which we can use further in our validation rules. So this predictor will populate the value for DPT given threshold data element. Period type, it should be same as my data set type. So if I am predicting the values for last 11 months, I'm calculating storing monthly values, then my period type also must be optimum monthly type. We should select organization unit level where we need to run the predictor. So if you run the predictor facility, it will take up the data for each facility and store that data in this data element for each facility. The same data element because it will be aggregated at the district, region, country level, so they'll be automatically adding up. We should avoid taking multiple levels here. Otherwise the predictor will predict value for facility as well as for region and district, which will duplicate the values. So I just want my predictor to predict the value for threshold at the facility level. Now this generator, so this generator tells me that what needs to be stored in that data element. So I want to, as we had seen this formula yesterday, I need to take the average of the doses given, average of the DPT doses given for the last few months and three times the standard deviation population of the same data element of the doses. So this is my formula, which I have selected in my generator. Then there is sequential sample count, which is what values do we want. So when I'm saying average, the average of the DPT doses given, I need the average for the last 12 months. This is a monthly predictor. So show me the average of the data elements for the last 12 months and sequential skip count means if I'm saying sequence is 12 months, do I want to skip any months in the sequence? Yes, I want to skip one month. So this will skip the last one and then calculate the values for the last 11 months in the predictor. The annual sample count zero because we just want the value for the last 12 months, last 11 months. If you just want the values for multiple years, the last two years or three years, you could say one. So one would consider one more year added to 12 help. So similarly, you could add on help for the multiple years. This was just an introduction to predict and it might feel complicated to cover. But if you have, if you're further interested in adding predictors or want to learn more about it, we have added the link for the same in the presentation that was shared yesterday. So in the user guide, you could just search for the predictors and you have all the information available there. How do we create our generators? How do we take previous periods? What is the meaning of each of these terms? What is a sampling period, a sequential sample count, skipping, annual sample count? All of these details are there with examples with these visualizations to see what periods you can consider. So anyone who wants to further look at the predictors and practice that, you could refer to the user guide. The link is shared in the presentation. But for our demo during this session, we will just use this data element where we are storing the values for the threshold. Whenever we create a predictor, it's important to run the predictor because when we run the predictor for a certain period, only then the values are stored in that predictor. Now, let's go back to the validation rule. And my validation rule is that I want to see DPT doses given during a month should either be less than or equal to the threshold value which is set in. So based on the last 11 months data, if threshold is n number, I should ideally be given vaccines equal to a less than n. I might be giving more vaccines which can be a case due to some XYZ reason. But then on an ideal situation, I should usually not cross my thresholds. And this will just show me as an alert on my data entry screen or when I run validations that when do I cross the threshold values for a dose admin store. So to add this validation rule, I'll click on the plus button. I will give a name to my validation, which is DPT. So DPT doses less than equal to. We can give the complete name of the vaccine. I'm just writing DPT to save time. And of course, I'll prefix with my initials to avoid duplication. DPT doses administers should be less than equal to DPT doses. I can add a description which will be available there. I can add an instruction that check with facility. I can give it again a medium importance period. I will be monthly here as well. The steps will be same as we've done before. My left side expression would be this is my validation. So my left side expression would be DPT doses administered for the total number of DPT doses given. So for that, I will go to left side expression DPT doses given and I will search for the DPT doses. Now I have three types of DPT doses which are given. So I could either add on for each less than one, more than one, less than and do the calculation and I could take the total step. Because for this validation, I might not be able to know which value is incorrect less than one or over one. And if I just show an alert for the total values, that might also be have same effect same use as I take individual values. So I could just say add up the doses given first dose given plus second dose given plus third dose given. And here you see dose one, two, three, I've added all of them and saved. Now my operator would be less than or equal because doses given should be less than threshold. You see this is still buffering. There is some issue here. Don't worry. If it's still buffering for you, the validation is still getting set. You could also create this rule otherwise DPT doses threshold should be greater than doses given. You can change swap the left and the right side if you want to do that. DPT doses given less than or equal to my threshold values. Now DPT doses threshold. I searched for the DPT here and we have this data element that we used for predicting the values. So I will use this data element DPT dose given threshold. I can just cross check by scrolling down. It's the same data element. I take my validation. Now I need to add this rule to a group. So I will go to validation rule group. I have a new validationization validations. Give a code description if I want to. I can search for my validation which is here DPT doses administer should be less than or equal to threshold. And I save this. Now what do I do? I need to run this validation. So I will go to data quality and run validation region again. Maybe few periods back and I will only run the validations for this validation group GA immunization validations and validate. So then there are a lot of values. If I can just search for one which has a threshold populated, let me just refine. So I had added this validation as never skipped. So even if the threshold is not yet populated, it runs through. But if I change it to skip if one value is missing, then these will not appear. Maybe I do some other page selection now. And the visitor shows that my validation is working correctly because I'm getting alerts for the validations. And it's running even when one of the values is zero. That's why I'm seeing so many values here. But there was just one validation rule in my group which seems to work correctly. Do we have any questions here? That's strange because I know there was some threshold I had run for April. But maybe there are multiple predictors created now. So it's just not running the one which I used. Yeah, do we have any questions or should we move to the exercises? So I see a comment on the chat. You will not be able to add or edit a predictor because we've not given you that access. But you can search for this predictor for threshold and use it in your validation rules. Since this session was focusing on the validation rule, so that's why we have not given you access for the predictors. But you could surely try that out on the demo instance of DHS2 and look at the managements which are there. I think we have a lot of time still left while we have one exercise pending for the day. Maybe we can take up questions if there are any or else I'll hand over to Pamod. Pamod, are you there? Yes, yes, of course. I think before moving on something different, is there any questions? Right. So we can handle them on Slack, Pamod, Okay, great. So I think all of you are waiting for one more thing from us, which is of course, like I saw a couple of comments in Slack as well. Let's chat asking for the word of the day. So let me share the word of the day. So word of the day today is consistency. So the word of the day today is consistency. You can see my screen, all symbols. So you have to open today's date, then under the word of the day, you have to write consistency.