NO CODE MEASURES using COPILOT in Power BI // Use AI To Write DAX // Beginners Guide to Power BI

Video Statistics and Information

Video
Captions Word Cloud
Reddit Comments
Captions
in this video we're going to go through how you can use co-pilots to generate Dax measures in powerbi we're going to go through how to enable it some useful prompts that you can use as well as some limitations and considerations to bear in mind if you want to use this all of that and more so without further Ado let's get started hi my name is fan and welcome to the solutions abroad YouTube channel bre over tips tricks and best practices when working with parbi I upload new videos every week so make sure you hit that subscribe button and the bell icon to get notified when a new one is out so co-pilot is a new feature that lets you use language models in powerbi to help users or developers explore their data copilot was a feature that was announced alongside the Microsoft Fabric and it lets you ask natural questions to your data to answer key questions give summaries and build entire reports all through these simple questions as of now not all of the co-pilot features are available yet in parbi So today we're going to cover just one of the features for co-pilot that is available for you to use now which is the quick measure suggestions which lets you generate Dax measures through simple natural language questions so here's a paria report that I prepared which just has a simple set of tables for the Northwind data sets we have some information here like the orders that were made how much they were the categories of the different products that are being sold when they're being sold and the calendar table which is something that we use for time intelligence purposes and a measure table which we use or we will use to kind of manage and organize our measures to start using co-pilot you will need to click quick measure from here or create a new measure using quick measure and you will see the options here suggestion with co-pilot now if this is grade out to you that's because you have to enable certain features in a couple different places for this to work so the first place that you would want to make sure it's turned on is in the option settings here in parb desktop you would go to options yeah so from the options settings here you'll need to go to preview feature and just make sure that quick measure suggestions is enabled you also want to make sure that the Q&A feature is enabled somewhere here because it needs to use a Q&A to be able to process the natural language questions that you put or use in your report once those two things are enabled you might have to enable some features as well on your tenants and that's because co-pilot uses AI or it processes it your data in a data center within the US and you need to give RBI the ability to export or take that out out of the geography where you're from if you're in the US this shouldn't be a problem however if you're outside of the US like me is based in the UK you have to explicitly enable this from your admin portal so you'll find your admin portal from your powerbi service here here I am in my parbi service tenants if you go to settings you'll go to admin portal this is where you'll find all the different settings that you can enable or disable we look for something called quick measure suggestions you just need to make sure that both of these are turned on so this one is enabled allow quick measure suggestions as well as this option allow user data to leave their geography so in this case if you're working with sensitive information and you want to comply with gdpr rules or if you have some restrictions towards sending or receiving your data from different geographical locations it's probably best not to use this feature now once all of those settings are enabled you should be able to use the suggestion questions with co-pilot from powerbi desktop over here so from here you will notice that it's almost pretty much the same as the quick measure suggestions however it seems to be powered by co-pilot so most likely in the back end it's also using some additional AI features to kind of guess what you are trying to ask so let's start by asking it some natural language questions and see what it returns so I'm going to type here give me the sum X of or order details multiplied by quantity so this is referring to order unit price multiplied by quantity and what I'm trying to do here is I'm trying to generates the measure that I typically create called sales which gets the total sales from all the orders within the north wind data sets by multiplying the unit price against quantity typically I'd write this as a DX myself however I want to see if co-pilot will be able to generate this easily for me without me having to write any Dax code but if you hit generate here so first of all what you will notice when I was typing and when I was finished typing co-pilot or powerbi understands the context of a few things that I write here so for example when I was saying unit price multiplied by quantity it autofilled and underlined it which means that it is referring to something in my data set when I hit generate at the bottom here it will will give you suggested measures as well as how it derived that answer the preview of that answer as well as the Dax codes that it would write so here we are it's showing me it looks like the right way to uh multiply the unit price and quantity to do the calculation it understood that I wanted a sumx of these values so if I want and I am happy with this Dax code I will simply just hit add and what it will do is it will just go to my my data model here it's called measure so we can just customize this a little bit and just call this sales and we're going to change this into a currency because we want this to be pounds and then we'll just drag it in here as a card and there you go so you have your sales which is what I would typically write but written for us using co-pilots now you can use these sales like you normally would with a normal measure so we can put it in a bar chart for example if you want to see the sales by different categories just do it like this so I'm just going to add some data labels here and I just and that's because I want to refer to this in a little bit I'm going to turn on data labels and I'm going to say for the values just give me like this so no decimal places and just the values so this total will be 1.3 million now let's go back to the co-pilot stuff and let's try to generate something else so I want sales let's say we want to add some more details to our our question so let's say we want to get the sales but only for a certain category or for a certain country now it takes a little bit of understanding of what data you have in your data model and how they're related but if you've got a pretty good data model to start with the suggestions that you will get here will be a lot more accurate to what you need so in this case I added and I enabled the label so that we can compare what the value of the measure will look like once we have generated it so in this case let's try to get the total sales of just a category so let's say Seafood so sales where category name is seafood so as you can see it recognized a few things here it understood what category name is it understand what sales we're referring to and what the seafood or what the category name is so if we hit generate here let's see if this will give us the right results so it looks like it's using my sales measure which is right and then it's doing the category name filter let's just hit show more let's see if that's correct 141,000 so it looks like it is correct so here we go so I've just generated that measure now so I'm just going to call the seafood Sals if we just drag it in here as a measure there we go so that should match up with the value that we have here in the label and verifying the results of the measures that you generate using co-pilot is super important just because it can give you anomalous results depending on how you set up your model or even the prompts that you give it so it's super important that you understand Dax or that you understand what it's writing and what it's returning before you start using it there are a bunch more other prompts that you can use so let's go through a few more examples so that you can see what other things are possible with the co-pilot stuff so let's say we want to get the average price of a certain category so let's say average unit price of condiments so it will give you for every single order where there are condiments it will just give you the unit price for that category which is really really interesting so you can see that it will give you a certain value here let's try to do another one maybe we want to get let's see the top three sales by category we hit generate you will see okay so it wasn't able to generate us anything so let's see it's probably because we need to be more specific with something here so top three sales by category name here we go so it's given me the first three which is the ones at the top here in my bar chart beverages dairy products and meat and poultry and that is exactly what I'm looking for and as you can see without writing any Dax code it's created a ranking for me all I had to do was ask simple questions here in this co-pilot box now that you know how to use co-pilots let's talk about some of the limitations and considerations to think about which we touched on a little bit earlier the first thing is all about data privacy as I mentioned before because co-pilot is powered by AI language models these models are and they live in US data centers which means that if you want your queries to be processed it needs to be sent to these data centers in the US and then spit the the information back to you now if you have requirements where you want to keep your data in house or in the same geography that you are working in this probably isn't the best solution so if you're working with sensitive data that requires you to keep your data within your geographical location it's probably best for you to first check with your business to make sure that you are able to do this with the data that you're working with the second as I mentioned earlier is to always validate your Dax so while co-pilot and quick measure suggestions give you a shortcut way to just generate Dax quickly is not a replacement for not knowing Dax at all and in fact it makes as develop course it makes my life a lot easier and faster just because I don't need to type and know all of the syntax that I need I understand the basic concepts of Dax and how it works so I know when for example a Dax measure that it's suggested isn't going to give me the results that I need so you still need to know this um and not always just rely on what the suggested measures uh give to you and that's really it for this video I hope you learned something and I hope you can Now understand and see how you can start using the suggestions with co-pilots in powerbi thanks for watching as usual give this video a like if you found it useful give it a dislike if you didn't so to do better for next time ask your questions in the comment section box below so I can help you and you can help others if you like this video we have a patreon page here you can support the channel and get exclusive perks like Early Access demo files and credits at the end of these videos thanks again for watching and see you in the next one bye-bye
Info
Channel: Solutions Abroad
Views: 3,257
Rating: undefined out of 5
Keywords: solutions abroad, power bi, powerbi, power bi tutorials, power bi for beginners, beginners guide to power bi, data analytics, dax, data modelling, data visualisation, business intelligence, how to power bi, power bi how to, power bi best practices, power bi tips and tricks, power bi standards, power bi patterns, power bi help, power bi tips, power bi 2023, copilot, power bi copilot, microsoft copilot, quick measure suggestions, ai dax, artificial intelligence
Id: UHEDJYbOy6A
Channel Id: undefined
Length: 12min 27sec (747 seconds)
Published: Mon Oct 16 2023
Related Videos
Note
Please note that this website is currently a work in progress! Lots of interesting data and statistics to come.