EVERY AI Feature you SHOULD KNOW in Power BI // Beginners Guide to Power BI in 2023

in this video we're going to cover all the AI features that you can use to level up your power bi reports we're going to go through things like being able to create sentiment score based on your free texts in power query using AI visuals like key influencers to understand the trends of your data to other features like looking for quick insights to speed up your data analysis all of that and more so without further Ado let's get started hi my name is fernan and welcome to the solutions abroad YouTube channel or recovery tips tricks and best practices when working with power bi 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 AI or using AI at the moment is a pretty hot topic with all of these different AI tools like chat gbt to make your life easier but today what we're going to do is we're going to cover all the AI or AI like features that you can natively use in power bi to either make your development life easier or to make your data analysis even faster before we get started in looking at these AI features I just want to preface this to make sure that if you're using and utilizing these tools make sure that you're still double checking the data and the results that you're getting from these because they're not always foolproof so it might speed up your analysis but you're not quite sure or you might stumble upon some issues if you're kind of of double checking the data so make sure you're double checking the data that the AI returns to you so let's start with a simple one so here we are in power query at the moment we have loaded some information about product reviews here in this table and let's say you wanted to create a new column that extracts the review rating from this text so you can see it says 3.0 out of the five stars and let's say we just want to get the 3.0 as a different column now what you'd normally do is to split this column or write some power query to extract just that first three text but an easier way and a more say more convenient way to do this is by using the add columns through examples so under add column ribbon up here you can select column from examples and then you'll see that it will create a column here and from here you can type what you want to extract from your table as you notice when I type 3.0 it started to fill up it looks for patterns in your data to see what you're trying to get so in this case we wanted 3.0 and this AI automatically kind of understood what the context of that is so you can see that it's filled up all of those different rows you can keep typing here to make sure that to correct and give more samples where data for the AI to work with it will also create the power query formula that you need so you don't even need to know what to write it will just write it for you so if you just hit OK you'll see that with just those two clicks you're able to basically create your own power query formula without having to type it yourself next are the AI insights that are available here in power query and these ones are sort of pretty interesting you'll find the AI insights under here under the home ribbon and you have a few options here to choose from you have text analytics which is a group of if we just insert here let's just have a look at what these different functions are and I did cover them separately in another video in Greater detail so if you wanted to know how and get a summary of all of these I'll leave a link to that video in the description box below but anyway let's go through it quickly here so as you can see we have three different text analysis that you can use you have the tech language which if you fed it let's say free text fields or columns if you have different languages in those rows it will give you an ISO code giving you or telling you what that language is which this ISO code then you can use for these two other text analytics functions to extract key phrases extracts key phrases from the free text that you have so it's good if you're for example like trying to categorize comments or free text Fields where there's too many words or phrases that are being used it's a good way to to kind of create sort of tags so that you can categorize them within your reports and finally you have this course sentiment which it uses AI or it uses cognitive services to look for key phrases or keywords that are being used within those sentences within those phrases and it will generate a score based of 100 so this is a good way to kind of understand in bulk what the sentiment is for certain let's say product reviews for example without needing to rely so much on the star ratings so here I've shown and I've kind of used this these functions in this Amazon reviews report here so as you can see the detect language has just simply we've just faded the review text here and it's given us English so that's what the detect language does you have the score sentiment which if you wait for a second so the score sentiment as you can see it gives a rating out of 100 so this is in decimals at the moment but 0.5 is 50 and obviously the higher it is the more positive the sentiment is and then finally you have the key phrases so out of this review text you can see that we have a few you know there could be a lot of words or phrases that are being said there and what this does is it generates a list of key phrases which you can now use to say and it's a good way to analyze for example you know what are for those that are low rated or have low sentiments what are the key phrases that they're saying so this is a good way to analyze that apart from text analytics you also have access to Vision which is there's only one function here that you can use at the moment which is image tagging so what it does is if you have a list of table with image URLs it will scan those images and it will look for items or elements within that image and generate tags for you based on that so it's good if you're trying to kind of categorize images creating some descriptions or some way to group them without having to manually create them yourself so as an example here you can see here we have a list of images here just random images of some gaming setups that I found online and they have image URLs here and if you wanted to see how that looks like it's basically this it's these These are the images and if you go back to here and you use that AI function what it will do is it will generate different tags like this so that you can categorize let's say what it finds within your image so does it see the stage does it see a monitor does it see a mouse does it see a table which is a good way like kind of like key phrases to categorize your data so these AI insights both the text analytics and the vision both require premium for you to use so if you only have Pro or even free you won't be able to utilize them just because you send that data to cognitive services online and you just get the result back and so you won't be able to use it if you don't have a premium capacity Azure machine learning though if you have your own ml model that you can connect to this doesn't need the premium so if you have your own model that you've trained within your organization you can just simply connect to it using this option so now let's go back to the report here as you can see I've already imported a few values here a few tables and a few charts as you can see we have a line chart here and then a bar chart a few AI powered visuals that you can use is if you have a look at this and let's say we wanted to know why this increase happened here if you right click here you'll see that you'll have this option analyze and if you click explain the increase what it will do is it will scan through your whole data set and it will generate a list of you know different factors that could have affected or possible explanations as to why this increase happened so you'll see here we have different visuals like this not just this you'll see that there are a few more different visuals here to understand okay what is contributing to this peak so for example sum of cells here you might want to show it as as a waterfall chart or you might want to show it as a ribbon chart to show ranking month on month which is pretty handy and if you like one you can just simply add this to your page and that's basically your Visual and the kind of analysis already made for you and you didn't have to do anything apart from just right clicking right another way is obviously at the moment we right click the to explain the increase you also have consecutively a way to right click and explain the decrease which does the sort of opposite way if you don't have a kind of time based chart can if you have a look at this bar chart for example if you wanted to understand beverages which is the highest selling category that we have you can right click that and you'll have analyze to understand the distribution of that and what affected that that sort of Spike and increase so you'll see here we have two different axes you have the the gray area chart which is the sum of sales and then in proportion to that next to it is the sum of cells for cells representative which kind of gives you a kind of way to show those proportions and where it increases and where it doesn't you can disable that like this and like before if you like this analysis you can simply just hit add to page and it will create that analysis and add it to your page for you along with this what you can have are what we call Smart narratives which is a way to generate sort of natural language analysis on the data that you have automatically with just a click of a button so I'm gonna create a copy of our line chart here just so that we have it here and then what we're going to do is when we click this chart we'll just simply convert this into a smart narrative which if you select it will generate some analysis trying to analyze you know that change in data so understanding when it increased or decreased and you know what are some different anomalies that it found and as you can see it's in natural language format so you can add it as part of your analysis in your reports or you can add it as a tooltip so if I just delete this if I'm able to delete this at some point thank you very much I'm just gonna drag this here so under format your visual you'll see under General it's not very easy to find under header icons and then icons you'll find the smart narrative tooltip here if you turn that on it will add another tooltip there on the top right which if you click it will generate that analysis for you without removing the visual altogether so that's a good way to kind of add that AI element in your visual without having to do any extra work speaking of natural language there's also the Q a feature which again uses natural language or you can ask your questions in natural language which power bi will then return results for you based on that so the Q a function or the Q a visual you will find somewhere here so this one q a so if we look for let's say highest sales by category category name something like this if you click that you will see that it will give you the the highest selling category as a visual right here and all I had to do was type the question that I wanted so fairly easy key influencers is a good AI visual that you can use to explain the distribution or the cause of values increasing or decreasing so you can find the key insights key influencers visual here so if you click it and let's say we want to analyze the increase or decrease of sales and let's try to understand it based on certain things like we want to see based on product name or based on category name perhaps based on order date let's add a few more things here and let's say we want to add customer country yeah so here we go so for example here you can see if we change info what influences sales to decrease so you'll see it will based on what we put on the explained byte we'll try to understand the correlation with the average sales and what is contributing to this decrease which is just giving us a different kind of list here from countries customer countries to different products and you can also see a segmented view of those kind of averages based on the total count which again is something that can speed up your data analysis if you're looking for sort of key influencers for a data's increase or decrease the decomposition tree is a good AI visual to allow you to drill down into kind of different dimensions of your data to understand what's contributing the highest or the lowest and it's pretty interactive so let me show you how that works so I'm going to make some space here actually we'll yeah we'll delete this one and we'll add decomposition tree like this and we're gonna add sales again in our analyze well here and let's say we want to understand uh category so we're going to add look for category name if you click plus here it will allow you to choose how you want to split your data so not necessarily all of your categories but which categories are let's say the highest value so you will see um it will now get us this sort of top three of those categories and then if you wanted to drill down even further let's say you wanted to see out of those beverages I want to know which countries those customers are from so let's I've just added country here in are explained by if you hit plus here it will allow you to then see higher low value and it will give you like a breakdown of those countries and you can add more and more levels here which allows you to kind of drill down using the decomposition tree I think I covered this in a separate video so I'll leave a link to that video as well as for my other AI visual videos in the description box below and normally detection is a good way to find out any data points that are going beyond the typical range that you set and it's a pretty easy one to enable so here for example example we have a list of sales or the sum of sales for different dates for the duration of what we have here in our data and to add this anomalies analysis here you just need to click this Analyze This Visual and then toggle find anomalies so you can adjust the sensitivity of this anomalies finder and then you can increase or decrease and based on this sensitivity it will increase more points or remove points here if you click on any of these points what it will do is it will try to create some or generate some analysis based on that how far off is it from your expected range and if you add further values in the explained by like for example you want to know based on customers country based on product or based on category name if you hit apply those possible explanations at the bottom will update accordingly now I'm not sure why it's not updating but if you add a few more I'm sure that will update too so maybe we'll let's see if we can select a different one so here for example yeah so you can see there is a little bit of an analysis here as well as some possible explanations um for this anomaly So based on categories or quantity sold or you know based on Country Auto clustering lets you group your data and segment them into kind of different groups so that you can analyze them sort of individually so here we have a scatter graph scatter plot Visual and what we have is basically each dot would be a sale and the y-axis that we have is the quantity for that order then the sum of sales is on the x-axis so if you wanted to kind of create some cluster groups here you can simply right click or go down here and select automatically find clusters you can set how many clusters you want or you can just use Auto let's just say we want to group this into four different clusters if you hit OK whatever is it will create a grouping for us you know based on proximity and you'll see that you have now four different clusters to choose from it adds this this grouping here which if you now add as a as a slicer if you wanted to analyze let's say just cluster one there we go cluster 2 cluster 3 and 4 and you know in total so if you wanted to analyze them and group them in this kind of logical order now let's move on to the service because there are a few things that you can use in the service to you know which is AI based that can help you get insights now the first thing that you can do is get quick insights from the power bi service from your data sets so here I have opened my workspace my premium sandbox workspace and I have just one report here a report and a data set and the report already you know generates some visuals that I've created for this data set but let's say you've just imported the data you've created the model and you wanted to get some quick insights on this um how you would do it is by clicking the Ellipsis icon and then get insight so now it's view insights because I've already generated it so if you click view insights so these are kind of quick insights that power bi is generally waited for you yeah so these are the quick insights that power bi has automatically created for us so you can see that we have some analysis here like how many comments when they've happened it's quite a lot of really good information here actually and it allows you to kind of before you even dive into the data you know look for interesting insights within your data and all I had to do was click one button from the service so this is pretty handy the last thing that I wanted to show you is this get insights from your reports within the service so it's this button here on ribbon and what it does is it will analyze and try to look for key insights that you have within your report page so at the moment here we don't have a lot but I'm gonna move to the history page here and if you click get insights right now what it will do is it will generate some insights based on the kind of visuals that we have So based on the the review volumes based on you know three month averages so as I hover over the insights it will highlight those parts of the reports and it will give some sort of analysis and you know some some natural language analysis on these visuals so pretty handy if you have a bunch of visuals that you've just chucked into a page and you wanted to get some analysis out of it and that's really it for this video it's quite lengthy because there's quite a lot of AI or AI like functions that you can use in power bi so I really hope that I haven't missed anything if you think I've missed anything leave it in the comment section box below so everyone knows it is available thanks for watching as usual give this video a like if you found it useful give it a dislike if you didn't wanted 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 really like this video we have a patreon page where 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

Channel: Solutions Abroad

Views: 6,074

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, power bi ai, power bi artificial intelligence, power bi chatgpt, power bi ai insights, power bi nlp, power bi anomaly

Id: 5URcVnV29Is

Length: 19min 53sec (1193 seconds)

Published: Mon May 01 2023