Power BI visualization best practices by Marco Russo.

good morning good afternoon good evening the my name is Charles Stirling chatting with me for last couple minutes is Marco Russo calling in from I didn't even ask him I'm assuming he's calling from Italy I should have asked about man Italy epoch is such a roll trot a visitor that who knows where he's actually from as will have been on these pals before me and Marco can talk with each other but you guys can't actually um or talk to us what I'm going to need you guys to do is chat in the chat window and let me know if you have any questions and Marco is going to go ahead and get started and she's going to tell me when is a good time to actually start breaking for those questions oh oh go ahead Marco what we gonna say now Jenna just for the question we didn't made an agreement but I would prefer I mean I mean if there are a related question we can you can stop in at any moment but I will try to keep 10-15 minutes at the end for all the Q&A but the in case there is a very very specific question you can you can stop me but otherwise we can have time at the end perfect and you're going to be out for the data inside summit and I this is to kick off a new training like you guys have got going on how long's this training been going on tell us a little bit about it tell us about your your journey out here and if I can get you to in case anybody doesn't know whom are Cairo so tell us about yourself and I'm going to hand the microphone to you my friend and mute myself okay thank you very very much of my name is Marco Russo and today we will see at this presentation about how to improve your dashboard in power bi I wrote a blog post today in my blog to explain now how we created this training and actually this time I this is a training I didn't create actually I'm the standard the speaker I I present this content but the entire content has been designed and created by daniele apparently who is the guy who works with us and you know take care of so many things about our website under the custom visual that many of you may have already used and the idea was to provide a set of best practices that anyone can apply to his own dashboards I have a certain experience with power bi especially for the data modeling for the tax language and I also very very long experience with the BI word but the designer dashboard is a skill that from a certain point of view requires some skill about the design but what we try to do is to provided the suggestion that can enable anyone to improve their dashboard at least avoiding the most common mistakes and of course trying to produce something that is also beautiful and not feel not only useful so first of all we have to define what what what is a dashboard and bi dashboard because many of you know power bi we don't we're not we're not referring to the dashboard of power bi the one that you have in the power bi service that is a set of ties that you can grab from different reports we are referencing to the more generic definition of a dashboard which this is the on the slide you can see the Stephen few definition which is a very good one in our opinion what the dashboard is a visual display of the most important information needed to achieve one of more objectives consolidated and rated in a single screen so that the information can be monitored at the glance so if you see there are a few concepts here that make a very big difference between a dashboard and a report single screen monitor at a glance and goal and the target is the user the business user so the idea is that the dashboard is something that you use a relatively often to get the very very quick analysis of the data and then you can go deeper with other tools with more rapport with other other dashboards maybe more detailed but this definition helps us in understanding what we we have to try to do with the - but we have to try to obtain with a dashboard and you can imagine when we say single screen what do you think no scrollbar no we don't have to you know go up and down then of course if you have a Windows File I mean if you have a mobile phone then you need probably to scroll something because otherwise it would be too small but when you see this in a in a desktop screen it should be a single a single screen so these are the generic this is the generic definition that that we can we can use and probably is that you probably have seen many many dashboards and water bottom what are the thing that makes a good dashboard and the mistakes that creates a bad one let's start to get a dashboard vector clearly has a number of issues I mean this dashboard is in a single screen the material that you see is material that we found in Internet so it's public and these are public - ones that we found and what what is that the issue that you see here well first of all we are not using the space the available space in a very useful way at the end of the day we want to see numbers and the numbers here are very small sometime are confusing sometime you see a lot of colors that don't convey any particular information and the choice of the color could be not particularly useful and in general this is a set of of also KPI for example the the gauges you see on the top on the top left just to display a single number it uses a very very big area of the screen so one a side effect of this is that the the number of information that you can provide in this dashboard is very very small at the end of the day you see that that there are just the four indicators and another table that has a lot of numbers in a very very small area of the screen so actually this is not very readable I think that also you are not able to read the content yes in my biggest screen in a very high resolution you can read the text that author of the table but actually it is not been easy to use so I mean okay this is easy there are a lot of effects that drive it it says it doesn't look so nice okay so let's start to see another dashboard at first sight we may say oh this is a nice dashboard right so let's take a look if you see it is a dashboard with some simple data that new reality it breaks a number of rules I will show you later that we defined 15 rules that we can apply to any dashboard in order to improve the dashboard and to get a good looking one and a useful one and this dashboard at UC violates six eight of these rules of these fifteen rules and we can see one by one this roots that are violated for example the first one is that we need to keep everything at a glance I will describe these rules in a minute but for example this this dashboard does not fit in a single screen and as you see we have a scroll bar and the scroll bar should not be used in a dashboard because it should be displayed in a single screen without having to scroll we have to keep it simple which means that we don't have to add in the in the in the report an excess unnecessary lines border or you know any a decoration that is not really useful moreover this dashboard is using a black background which is another better idea and I will explain this later we have to keep the elements aligned and as you see this is the just issue if you want this just a you know aesthetic problem but a dashboard that doesn't look good doesn't convey a good message so we should align everything in a good way then there is another problem that does if you look at the symbols use it here there are a number of names and and symbol that do not are not self-explanatory and there is no legend explaining what is the meaning of the symbols that you see sometimes dashboard has to be shown to a business user that don't know exactly the acronyms used in the business simply imagine when a dashboard is used in a in a meeting with with a stakeholder or shareholder and and they maybe they they don't know all the details of the business so having some explanation of what you are looking at that is important we have to start from zero and there's another rule that is violated by that at least a couple of gauges that we have here violation of this rule makes a few matrix to appear similar whereas the real position is very different if you have another metric so actually we have gauges that are next to each other but actually you cannot compare them you cannot use them side by side we have to reduce some time the number we have the shorter number because a simple level of detail of the simple metric is not very useful actually looking at all the digits looking at the decimal point some time is not used some time I want to see 1 million 2 million 100k is enough to understand the size of a metric that we want to display we have to choose the right color and the this is a very very underestimated problem but actually when you choose color you have to think about the fact that not all the people you see the color in the same way there are a simple number of people and actually five to ten percent of the population may have some different way to to to is it I mean is affected by colorblindness but sometimes it is just a different way they recognize color and so if you use the color to convey an information you have to make sure that they are able to recognize the color and for example in this line you see that people affected by frightened appear they see these instead of I go back one slide instead of this so in this slide that we are using the green color in the engage but this color is actually looks very similar to the color used in the in the bar chart that you see at the bottom so this is another thing that you have to be careful about and finally pick the right chart probably everybody knows that everybody knows that that you have to make a choice between all the available visuals that you have in power bi or in any other tool and you have to make a right choice and actually all the highlighted charts probably are not the right choice for example the bar chart is not the correct choice when you want to show something that is a trend over time a line chart or a sparkline that is a different kind of flying chart would be the best choice in this case what is the gauge is usually not a good choice because it requires a lot of space and that we may have had a better visualization by using both a chart for example because the bullet chart will type providing that providing exactly the same information in a much more condensed of space in a more clear way especially if you want to compare them together so this is just a list of the thing that we can fix in this in this dashboard so let me let introduce let me go to the next slide sorry let me introduce you the system rules that we should follow to design a good dashboard so okay a perfect dashboard but we try to do let's say the good enough it would be a good start so these are the cysteine rules that we coded in the train in the course that we created so what I'm showing you today is an overview of all of these rules and I would spend time on one of those just to explain you what is the process that is behind each of these rules so the rules are relatively simple one by one these are very very simple steps that you can follow the first rule is that that you have to design for a target which means that you need that to define the goal of the dashboard what is that the target of your dashboard the second is that you have to keep everything at a glance which means that for example you don't have to scroll up and down to see all the data that you have in your dashboard the third is that keep it simple this is the rule that we will spend more time later to analyze in more detail but in a few words it means sell less decoration avoid a picture background or a black background because they this elements don't add any value to your information to your dashboard and actually could be a distraction the 40s align elements and I mean this is easy to realize if you see charts and text editor in a good order it is better the fifties that be consistent so don't choose different colors and charts just because you want to use all of them that's not a good reason if you have the same chart type across the dashboard is that the type that if that chart type is the right one the dashboard will not look boring if we look professional so don't confuse the the faculty just because you have a tool you don't have to use that tool you need a reason to use it the six is a highlight the most relevant information which means that the not all the numbers are the same the dashboard resolved with some data that is more important than something else so it deserves a larger arrow of the screen a larger form the position matters there are parts in the dashboard that are more important than others because we as humans are used to take to to put more importance to an information that is put in a sitting area of the screen or of a report and the seventh is be clear which means that that we have to explain what we are showing and some time using an acronym can be complex if the user doesn't know what could be the the meaning so having a legend sometime is really necessary the ATS start from zero which means that sift and chart but not all not almost all the charts just simple charts requires you to start from the zero for the axis has to start from zero others don't have this need so we can see when and why to use this rule the 90 is the shorter the numbers as I said that we may have to reduce the length of the number of digits that we include in each metric that we display as number and not just as a chart because some time having 1 million is better than writing one followed by six zeros sometime it is is just good enough having this kind of visualization the tenth is a show the context data should be compared with something else sometime you put a number that's fine but you usually want to compare this number with the previous year with a budget so the context is how I can compare this number sometime this is a in form of a chart sometime this is in form of another number that you may display close to the other may be smaller the bit with a different color if this makes sense and is consistent with your with your dashboard xi choose the right color we already already mentioned this before there are people who don't see the same colors and don't see colors in the same way as you do so you may have to figure out whether what you're doing can be ended dude by anyone that we will see your dashboard the twice is and this is a very very common mistake I had to say design dashboards not report I mean there is nothing wrong using power bi to create a report but actually power bi is a tool that has been designed to create dashboards and to create what we call reporting power bi is actually a dashboard when we told the general term report is a something like a paginating reporter that you creating reporting services a long list of names and values and Sun in a dashboard we don't want to show all of the customers usually we show the top of five at top ten we show something that is visible in a single screen then sometimes sometimes but as an exception we may show a longer list with a scrollbar because we are not able to show all of them but this should be the exception of the rule having only listed that have a scrollbar and if you feel the screen of of this language that you are not creating a good dashboard you are really something that it is not providing a very very quick and easy to understand information the 13th the 13th roll rule is a show variations and if a difference is important choose the right way to display it and some time though if you want to highlight how the data were rated over time you have to figure out a different way and also different way to measure this difference and to display this dispatch rather than adjust the show into number in a way we there are many many different ways of different results depending on what you want to highlight the 14 is a leave the noise off which means that you should not suggest relationships that don't exist it is very important sometimes just putting two numbers to chart close each other may suggest that there is a relationship between unrelated the matrix if you know that this matters have not really together just having a some separation some space in the middle to be important to define an a separation in different parts of the same dashboard and a-15 rule is of course the simplest to explain to the wrapper be the right chart so use the right tool the right the visualization to display what you want to provide to the user so at this point let's take a look at that okay marker what can I do when I follow these rules and I will show you in the next slide what is the before and after of this and before is on the left side of the screen we have a dashboard it actually is not so bad but in reality does not that doesn't not have it has a number of issues three it violates the fifteen rules the dashboard on the left and the dashboard on the right provides the same information in a more professional and clearer way and what you should think about is what is the dashboard that I would like I would like to see every day is the one on the left or is the one on the right because I know that many of you think Bettina in a meeting the first time the one on the left you know more investors because if you have a picture actually when you can see this data all the day and you have to navigate you know there is a difference and I want to show you this by oh there is an issue with the screen yeah Marco you could get you to reshare your screen I think that might help we just lost it about a mil Gaya so it hasn't been gone very long um okay and Carla hi welcome I said that you just joined you only missed about 21 minutes okay so let me let me go wait a second so I have two oh two two two two nothing be easy to hang out there okay okay so share screen make sure it's it I think it's coming through and while you're straight we screen assuring I'm happy birthday to Michael Michaels calling in from Israel and anointing us on his birthday how cool is that okay okay not not yet okay now milk it I see you talking yes I know I'm talking but I because I see in the actually there we go there we go I hit the screen okay networking now the patience okay good you're gonna see it a couple seconds so just wait wait for it I see it on the moderator screen so okay and we're live again Marco go ahead thanks okay sorry for for the problem so I was I was a saying that the the to dashboard that you see in the screen I I hope that the resolution would improve in a few seconds on the last time you see the the dashboard that violates the fifteen rules and actually it it is not so bad I was saying but actually it could be improved and the one on the right is the is the same - for the same content after the cure after we we improved the the dashboard by applying all the rules that we have now I will try very dangerous operation so I hope that this will work I now opening the real dashboard because this is the power bi 5 that contains the dashboard violating all the rules and if this is visible I hope that you will see that once I enlarge everything the problem is that this dashboard requires a scroll bar to go up and down and especially because you are probably seeing this in the you know in the streaming probably the experience is not very good because when I scroll down you have probably to wait some time before we actually scroll the screen and by the way this is a real dashboard so we can actually click on a different visor and we are actually changing everything so the problem is that if I want to show everything I end up doing this so I end up and doing this but this reduces everything so everything is smaller I'm wasting probably almost half of my screen just for nothing so this is not a good idea now let me show you the same dashboard after the improvement and now this one this one here we go so you see here that this is the same dashboard actually it is not the same is the the improve it one and this - border actually has what is the main difference that you see you see more numbers more information more data in an easier way to be read and used so actually the final goal of this tool is to provide information for the business and this information is better available to the user if we avoid a number of unnecessary addiction like graphic larger fonts logos that are not really useable this is just an example that I will explore now in more detail so the idea is that I personally will prefer to use this dashboard rather than the first one when I click something I clearly see all the changes I can better appreciate what is changing every time and every monitor of my on my store so this is a the difference between the two so let me go back to the slides now so that I can all let the second well I lost my slide here we go okay and let me go back to the rules and the rule that I want to explain you in detail so we can go step-by-step in okay what we have to do for each of this role okay let's let's examine one it should be simple so keep it simple so what is this rule about and the foundation of the design is simplicity and so the idea is that we want to simply remove visual effects border gridlines images we don't want a black background that usually a white or white background is much much better think about this the dashboard that has a white background can be printed if you have a black background how do you print the dashboard now I'm one of the guys that don't print anything I will never print a dashboard in my life I promise but I know a lot of users a lot of business users that want to print the dashboard for any reason and so actually this is something that you have to think about because if you print something the path of black background the result is usually horrible and finally avoid using too much typography too many different fonts bold italic keep it simple okay so when you want to you when you want to emphasize something you can use bold it should be an exception you can do this one maybe two times in a dashboard zero is always a better number so that's the idea so let's go in details okay we have seen we have quickly seen a dashboard that represents the data of a store we have the store we have says you know it's the classical canto so actually this we call this a PPI store just to give a different name we had to introduce some different concept but the basic idea we started from the information from the classical sample database composer so we have we sell goods we sell different products in different countries and actually we wanted to create a dashboard for the store and what are the unnecessary decoration in this in this dashboard first of all we are wasting 1/3 or at least 25% of the area for a picture and the picture is also you know it visible in the background where we have the white the white background but in reality in transparency we see part of the picture and this is distracting this is this is something that does not help in any way to read the numbers and because we wasted this amount of space we had to use a smaller font for the for all the other information look at the sides of the title why the size is so big look at the at the comment the text on the on the Left established in 1972 is one of the largest chain of the hyper marks in the US okay if I have to do a presentation in a conference I may want to create a slide like that a dashboard for my managers I have to remind them every day when this business started maybe it's not necessary information look at the numbers that we have on the right we have this bigger number which is maybe a good idea because these numbers are important but why we have to use a border around is it really necessary probably no and what about these two numbers the two numbers that you see in the bar chart you see in the bar chart we have two numbers let me see if I can I like this oh let's see highlighter this one so you see here this area here these two numbers why we need the dis to numbers because actually it doesn't help us in understanding the overall trend and it is really unreadable if we if I really need to show the highest value for a chart that is probably better to write this number in a separate way so sometime because it is so crowded and so small it is not helping in reading information so let me show you another another point which is a what we obtain after we remove this pointless graphics and as you see now we have our data now our data is starting to breath so we have we have more available space we we can separate we can use a larger phone we can separate each visual so we have we can put more data in our report so by just removing unnecessary information we created a more readable dashboard and actually it is not usually it is not I mean at the end of the day if with a clean design you can obtain also a nice-looking report so it's not rocket science to apply this rule but actually we can obtain good good results with these simple steps and one by one step by step we can improve the dashboard so another in from another interesting point is that you will notice that now we in some way because we have more space now we group the few charts and we separate the few charts from other I will not go in detail about the meaning of each element but the idea is that there are part of this dashboard that are not related to other parts of the same dashboard so we kept them separated and this space here here so you see here this area this area this is blank blank blank it's not the just because we didn't have anything to show we wanted to use space to create a separation but instead of writing a line to separate things with the white background clear space to keep things separated it is actually better usually better than trying to use other other approaches to this sorry so at this we can take a look at some other set of examples where we have something that that from a graphical point of view could be I would say a caching so you can actually oh it seems interesting it seems nice but actually there are in terms of usability the dashboards that I'm going to show you uh hides some some error that are in the same area of the error that we want to correct one to fix with a keep it simple principle and for example in this case what are the problems we have here actually this looks nice I mean from a graphical point of view it seems an elegant dashboard but what are the problems the first is that the chart the chart that you see here is not very easy to read what does it mean let me highlight this so this chart here all the yellow is not visible so let me use the laser pointer this chart here if you see this this chart is not very visible and I mean it's not cannot be understood understood very well because that we have areas that are overlapping and because of the transparency we are not hiding every anything but actually what's the meaning of this why we have an area and not just a line chart what is this point it is very hard to understand without the colors and what are the metrics that we are displaying moreover the choice of the colors has been a very bad one think about this we have the purple color here for this brand then the same purple color is here for the new customers the same purple color is used here for Wednesday and Saturday now if I see the same color I imagine that there is a relationship same color same metric or I mean it should be something related it seems that this brand is using is sold only on Wednesday and Saturday or this is the the percentage of the brand that not there not this is the customer so it's something else so it's the the choice of the color is a bad one so may be that the color choices are consistent for aesthetic but are completely inconsistent for the dashboard for the the message that we want to convey with with a dashboard so two mistakes color choice overlapping area in the chart in this slider again we have something data from a graphical point of view you see oh it's an elegant dashboard but what are the promise here I have done what is this chart because we feel the this area of the color is this a stack or a chart or not well if I look at the numbers here I would say no because if the color is greatest stays channel sales and direct sales as I see also from the categories here if I do the math I would say that the the number that I see here is not the sum of the other two or not I mean it's ambiguous I cannot realize what I'm looking at and the other problem is that they're here in the gauges that we have we don't have a reference when you see a person page you need to know a percentage of what 80% of what so eighty percent sixty four percent seasonally because the total is not 100 these numbers cannot be cumulative so there should be something that is a you know we have a number this number is a number is a ratio against something else is it the total sales maybe is not very clear so it is something that should be better described in the dashboard and as I said maybe the legend some time is a good idea in this example have a nice picture in the background there's a biker thing problem is that we are not adding any value and actually it could be distracting to see this road and this you know sunset the nice a nice crime but actually it's too dark to be really seen and it is not having any value and of course here we have the problem of hiding information in this case we are hiding okay there is a motor response here but we don't see what is happening under this area in this area where is the red one for example moreover we have here numbers that are really not readable why we don't use a thousand separator or maybe we short to the number using needles or or a hundred thousand for example so also in this case we have a some problem and last but not least what's the meaning of these colors I don't see what is the meaning of the color and by the way this color changes over time to these other color so I don't understand whether there is a relationship between this color as your green and maybe red and the colors that I see here also because this yellow or amber one is not displayed and it's anywhere else so I don't actually have any clue about what I'm looking at in this report the last example that I show you is about much better report actually in this case we have a fewer number of errors because this dashboard is easy to read has a light background but still has some unnecessary decoration but I would say that we have a much lower number of issues in this dashboard compared to the previous one in this case we have some issue in comparing the numbers provided in the tables and in gauges I would say that this is not very clear how can I read this yes I have the legend here but actually there will be a better visualization here so it's not a problem for the rule number three probably this is also involving the rule number fifteen and we can improve this we can actually provide the more details about the visit which frequency we could add a trend we could add a line of the trend over time for this data but as I said better can be improved but actually something much better that with what we have seen so far so one of the rule the time I mention is a try to avoid the dark backgrounds and just as a test we try to apply the dark background to our dashboard once we apply the the improvements that I described and is the result now if you look at this it doesn't look so bad unless you have to print it but if you look at the two dashboards the black and the white one side by side what is your preference and from many point of views I still prefer the white background so the white background makes it easy to print and it is much easier to avoid a number of trade-offs that you have to make once you start using a black background because usually most of the visual are also not designed to be displayed without put up with a black background okay so just to recap what we have seen in this rule is that a good dashboard should be simple and easy to read we have to try to remove unnecessary information so that we would have more space a bigger area and a bigger choice of chart so for the most important information and the dark background that I know many people like it but actually some time is not the best choice so at this point I will spend some time talking about the chart reference as I said that we spent time trying to classify the visual that we have in power bi even if as you will see this classification can be applied to any other product but of course we we went into Ted for power bi in particular but most of them most of the visuals are available so another another product in the idea is that we wanted to make it easier to make a decision about what is the right chart I have to use for a particular metric so we started creating a number of categories that we can use to group the visuals so these are the eight categories that we define so we have a comparison change over time but whole flow ranking spatial distribution and correlation the yes I know we may have had the probably other two or three or four categories but actually having too many categories brings you in the problem of okay which is the right category so actually when we have to display something it is really easy to recognize what you want to do in this list of categories usually you have maybe one or maybe two sometimes you have you may have the dot the dot about two categories because you may want to do two things together at that point consider creating two chart one for each category even if as we will see there could be visuals that belongs to two different categories at the same time but usually we try to classify each each visual preferably into a single one so once we know okay I I have my data I have my measures my metrics my information I want to show I want to compare the current citic at the face for the current here with a with a with a badge better maybe with a previous tier I want to see the trend over time I want to see how much percentage of each or the category we are selling a compared to the entire volume so it's relatively easy to identify the category so this is the first step the second step is okay I know that I have to pick a chart of visualization for this particular category what are my choices at this point you go here and we created them these are a visual reference this is a PDF version I want to show you this because it's very very nice this is a PDF fun very high-resolution completely vector format you can zoom in zoom out and you can print it in a very good quality if you want keep this PDF updated over time you see that this is updated for the May 2017 and you can download this for free so actually if you go to this URL Marco or you go my friend I'll give you just what happened just now over the first nine dhoka okay and we need to reshare your screen okay wait a second oh and while we're doing that um I actually I can see it right so we just got it back um since I've already interrupted you I'm gonna ask a one of the questions I think it was actually yeah yeah thank the Benjamin diamond I already know the answer but we're gonna actually get it on the audio track says are all the examples done in power bi it looks like you've got some from other tools and I already know the answer but I want to get in the auto attract no yes I explained in the bad examples we use the examples that we found on internet and certainly these examples have been created with other tools but the examples that we show for applying it the rules are all in power bi so especially the dashboard that you can see the first and after the cure after the applying the rules these are power bi and in fact I can show you hopefully we will not we would this will not consume too much bandwidth this dashboard is a power bi dashboard and one second these the final dashboard is also a power bi dashboard and actually you can click on one month and you see that these are real I mean it is working it is not a fake one then of course in order to show bed examples I we didn't use RBI because we found real examples I mean these are the all the dashboard that we have seen where examples or actually a real dashboard if we found on internet and and they were put public so we didn't stole anything so we actually use something that was published in public websites and just to show what could have been an example of a mistake that you may do even when you do a good looking dashboard that actually has some problem in I mean it could be improved I now say this completely wrong but just could be improved you have a time for one word you want to keep going okay um up to you yes well just one second one second so we can open the Q&A I almost finished what I wanted to say is that once I am in this visual reference that I have as a PDF you can download the PDF from this URL and let me show you because in this page actually we have a number of things so if you go to the if you followed the URL you will land in this page and if you scroll down just before the list of the videos and most many of them are also available for free but this link enables you to download the PDF so you see the position here you scroll down when you just before course curriculum you have this this link and you can download ok so once you download this PDF what you can do you can quickly figure out the choices that you have for each kind of category and and for each visual we provided a further classification we have with the yellow background the preferred choice with a white background something that is okay good but not the the fabric choice and in gray background what we do not recommend to use and of course there is an explanation for offer for each of them but I will I would spend the house red but I think this is even just getting this and maybe discuss or figure out the why this is better than other it could be something that helps you making the right choice so now let me see if I have okay so I think that we can we can we can open with the Q&A because I don't have anything else to show at the moment I mean there is a course online but this would take I mean we have six hours of videos in the course so it would take a long a long time so charts go on okay so zap is asking he often runs into the problem and I actually running this problem is as well is that his dashboards starts getting so much information that they actually start getting hard to follow at what point you have any guidance at what point to start breaking it up into two separate dashboards and zap I apologize if I'm not perfect this is a very good question is a very good question and the idea is the following first actually in the course we showed that the initial dashboard that I shown you in reality has a data from three different departments of the company so at that point of the first day if you have too many things and you want to split down the first thing is that and a dash for should be for one area of the company now it is true that the CEO of the company want to see everything at the same time when you want to see everything you doesn't mean that you want to see everything at the at any level you want to see everything at the top level so when you go at a very high level you you are at 10,000 feet and you want to display a few metrics then if you drill down and of course at this point you have another dashboard but imagine one is the CEO dashboard then the CFO has another one and in the CFO we provide more information and remember the dashboard is always a first approach then you always have reports behind if you want to go in detail you want to see the list of 10,000 customer and the sales for each customer and the revenues of each customer that's a report but the dashboard could show you the top 10 customers or to show you the amount of the cash flow the trend the the trend of the credits and of the forecast of the the cash flow so these are the information for the CFO and of the 20-30 metrics that could be interesting for the CFO the CEO may be interesting to four or five it was in the same dashboard he needs other four or five metrics from other areas of the company so yes sometimes you have to split the dashboard true and the criteria is that the first you have to realize what is the level of detail that you need the higher bill I mean the more you know the more aggregated is the data the more high level is the data the more you can include data from different parts of the company because probably who want to see the dashboard you want to get a very very broad overview the more you go deeper the more you go in a single department of the company the more you can go in in detail which doesn't mean that you have to provide the list but you can provide more metrics about that department sounds good yep yep um Phillip um my friend from the focus conference I'm not positive he's asked him and macey if you you can actually make sense of it and actually you you and him know each other um has anyone ever suggested using a dynamic dependent validation list to choose a visualization I think he's talking about yeah that reported that the data's easy like yeah I think that this is a good I mean it depends on that you you mean that I'm reading the the question now I would say this okay let me see if I interpret the question in power bi there could be some wizard that helps you choosing the visual through okay but this is a good suggestion for the features of a power bi if you mean in the dashboard the user should be able to change dynamically the the visualization I don't think it's a good idea I mean the user of a dashboard should consume the dashboard if you think about an user that interact with the dashboard I mean one thing is choosing one month and one thing is going is in the dashboard for doing advanced analytics if you I mean with power bi you can do many things we're not saying that power bi is here only to design dashboard the dashboard is one of the outcomes of a power bi report the a report you may have different goals if the goal is to design a dashboard that many people would see every day click they're taken over overall I mean take the big picture of the company and Sun that is one approach another approach is you have a PBX file with twenty pages and Paige has a number of Lizer and a single big pivot table two charts a number of visors because it is in reality a dynamic report that's fine just don't call it this is not a dashboard but it's a tool for doing analytics for doing a investigation over your data that's completely fine it is just a different goal we have to to figure out what is the goal if the goal is a dashboard the dashboard has a system role and has certain rules from the dashboard you can jump it to a report which is a detailed report where you explore the data when you navigate into the data but usually the ratio of users between a dashboard and our report is 10 to 1 or 100 to 1 because for 10 years our debt to look at the dashboard you have one user that say oh I want to see this number in more detail and he starts drilling down the data one ally to explore data to see who is the customer who is contributing for 80 percent of the revenues this month so this is another thing is not the dashboard that makes perfect sense player yeah yanno makes for fixes um Michael sharper I don't know if we watch my presentation Ares is also an advocate of some of the design patterns I use asks the claim dashboard looks nice but the visuals are not separated as much as you'd like isn't it better to put a slight background the visuals and use that background for separation and this is something I actually do all the time so where you had some background ok I understand that sometimes I mean let me do let me do a quick curve background the reason why we created custom visuals I mean a number of them is because we needed them for this course for example the ballot chart is the custom visual that the solve certain visualization that the defaulter visualization that you have in power bi weren't able to solve at the point that and I understand what is your point you say you find not able to obtain with the properties of the visual the kind of separation that they want or the kind of visualization that I want I could use the background as a way to obtain that feature I completely understand that okay and what can I say actually I will not use that as a separation but as a way to create a reference that you will not obtain in the chart it could be a good idea but we are doing that because of our current limitation of the product I mean it is a workaround it's not the way to go I mean if you really need to use it why not but then saying is that it shouldn't be considered that the way we design the dashboard actually we design dashboard when we have to design certain - when we use another tool like for example a completely graphical - to design the dashboard as you know as I was a very to get the delay out we we I mean Daniele who could do this he creates a design a completely artifact that is not a real dashboard it's just a design and starting from that design we try to understand how close we can go to that design using power bi and of course I understand that that sometime you have to do some trade off the game I hope that in one or two years we the need for doing this that workarounds will be much smaller than that so my point of view I will not start the designing a background that I will use to put the dashboard on top of it also because my experience is that once you have a new version of power bi sometime you have small changes that completely destroy your alignment to the background we have seen this happening so I don't want to go there I don't want to spend time to align of the few pixels every visual every time we have a new version Barbie I just because they changes some you know CSS and now there is a difference of the peaks in in this case I don't want to lose time for this all right that makes sense um I you have to actually if you get time you have to watch my webinar because I actually cheat quite a bit and avoid a lot of those pitfalls um so a couple the questions have been around getting access to your slides so yeah you have a link here um are the slides in that PDF or what's the best way to direct a good point let me think can we add the some content to this page because I think that people will visit this page later so the easiest way to it is to provide a link to this I mean otherwise someone has to watch a 60 minutes or two to get the the URL so actually the easiest way see if we can add some link or content in the comments here yeah we wasn't in the description I'll add this this link it's okay perfect okay in the slides the slides are this link very good so I would provide you the PDF up with the slides and we will see where to store it and we will provide you a link and I mean just after the the session we will fix it okay okay sounds perfect perfect and there's a couple more questions but I think what I'm going to do is we're going to give our our viewers we have 304 online so thank you very much for joining to give them two minutes back and I'm going to give the give it the microphone to you to go ahead and close and Marco really looking forward to seeing you next week and thanks again for doing this webinar we really do appreciate it yeah thank you very very much Joseph for this opportunity to present this training as I said we launched this course just yesterday so we're very proud it was a very very longer longer work so this is the preview of the power bi dashboard design course you can see this at secret bi sequel dot VI is less dashboard is a short URL we created just for this I hope you will enjoy remember you can download the PDF and give us feedback at tell us what you think about our design print suppose and thank you very very much for attending thank you very much and I'm gonna go ahead and drop your clothes you guys have a great day Thanks

Channel: chassMSFT

Views: 127,956

Keywords: PowerBi, Hangouts On Air, #hangoutsonair, Hangouts On Air, #hoa, #hangoutsonair, Hangouts On Air, #hoa, #hangoutsonair, Hangouts On Air, #hoa

Id: -tdkUYrzrio

Length: 59min 41sec (3581 seconds)

Published: Wed May 31 2017