Power BI Workflow

Quick reference

• 00:04 Just before we get stuck into actually building stuff,
• 00:07 I want to talk about the dashboard development process.
• 00:11 Whether you're in Excel or power BI, this is virtually identical and
• 00:15 the key thing is that it's a five-step process in order to build a successful
• 00:19 business intelligence project.
• 00:21 The first place we start is where we get and transform our data.
• 00:25 This is a big challenge obviously because not all data is set up curated and
• 00:30 cleaned and ready to go.
• 00:31 As a matter of fact, this is usually one of the biggest problems that happens for
• 00:34 analysts is trying to get clean data to work with.
• 00:37 Power BI has a very very big set of tools available in this area and
• 00:42 it uses a technology on the back end which unfortunately you won't
• 00:44 see tagged inside the product.
• 00:46 But it's called Power Query, Power Query is a software that
• 00:50 is virtually identical between Power BI and Excel.
• 00:54 And we're gonna get some time into working with this to show you some of the great
• 00:57 features set here.
• 00:58 Now this course isn't gonna go super deep into this particular area, so
• 01:03 it would behoove you to get and
• 01:05 take a course on Power Query as well to really master your data.
• 01:08 But it is a very, very useful tool I'm gonna say that tool sets
• 01:13 are portable between Excel and Power BI quite seamlessly.
• 01:17 Now once we have some data, the next thing that we do is we go and
• 01:21 we relate our tables together.
• 01:22 So we'll take multiple tables and
• 01:24 we'll relate them together into what's called a model.
• 01:27 This is also a complex topic this uses a technology called Power Pivot
• 01:31 which again is synonymous with Excel.
• 01:34 And we're gonna take a look at how to do this, how to actually work through some of
• 01:38 the relationship problems and whatnot as we go through and build up on this course.
• 01:42 The third thing we do is we go and develop what we call measures.
• 01:47 Measures are portable formulas and they're very similar to the Excel's syntax,
• 01:52 as a matter of fact, if you spent any time working with Power Pivot in Excel,
• 01:56 these are power pivot formulas.
• 01:58 It is the same technology layer there.
• 02:00 Again we are going to get stuck into looking at some of these things but
• 02:04 we're not going to go ridiculously deep in this area either.
• 02:06 This is another area where if you really want to master this technology layer it is
• 02:11 so big that you should really take a course on specifically on
• 02:14 Power Pivot to really master the language called DAX which drives the formulas.
• 02:19 But we'll give you enough to get you started and
• 02:21 get you running quite some different way here.
• 02:24 The next step the fourth step is where we go and develop visuals.
• 02:28 And we're gonna spend a lot of time in this area here because this is what
• 02:30 makes Power BI sing.
• 02:31 It's all about the visual, it's all about the dashboarding,
• 02:34 it's all about making your data look pretty and conveying messages,
• 02:37 so this one we are going to definitely spend a lot of time on.
• 02:40 But first we obviously have to do the other things, where we get into our data,
• 02:43 we relate everything together and we develop the measures that we want to see.
• 02:46 So these things all sort of work in a big circle.
• 02:49 Once we have developed our visuals then we publish, we explore it, and
• 02:53 we share our stuff.
• 02:55 But then something happens,
• 02:57 once we do that we realize that the process goes around.
• 03:00 Because as soon as you start exploring your data once you've published it and
• 03:03 you share it with other people and they start exploring it, they come back and
• 03:06 ask for more.
• 03:07 And at that point we're going back to the well to go and get more data, and
• 03:11 transform it and whatnot.
• 03:13 So at this point you think well okay that makes sense,
• 03:15 this process is completely cyclical.
• 03:18 And that's true except that at any time we can go back and do a previous step two.
• 03:23 This is the way we build business intelligence,
• 03:26 is we iterate through things.
• 03:27 So we might get to the develop visual stage and realize,
• 03:30 that hey, you know what I forgot to develop a measure.
• 03:32 Or maybe I need to go and
• 03:33 transfer more data because I am missing something that I need for my visual.
• 03:37 So, the process is linear but it comes with these different regressions
• 03:42 that end up going backwards and forwards all the time in order to cycle through and
• 03:46 get our data, so, this is something to be aware of.
• 03:48 Generally, the process is linear but it does loop back on itself many, many times.
• 03:52 So, don't expect to get it all right the first time.
• 03:55 It's a really important thing here, you don't wanna be beating up your
• 03:58 ego over this because you have to keep going back and
• 04:00 redoing stuff that's part of the natural development process.
• 04:04 Now, the other thing that I wanna just throw out there is you
• 04:06 really want to try as much as possible to keep your business intelligence targeted.
• 04:11 And this is really difficult, we have this thing in programming called scope creep,
• 04:15 and this happens in business intelligence development as well.
• 04:19 You'll put your file out in front of someone and say, hey,
• 04:21 that's great but could you just add this whole other department into this
• 04:24 analysis that you did?
• 04:25 And then, yeah, I can do that and you run back and do that.
• 04:27 And pretty soon they want every department and they want the competitive analysis vs
• 04:32 the competition, and they want to compare it to how things are going in the economy,
• 04:36 and they want to compare it to whatever the last foreign policy change was that
• 04:39 somebody made in one of their governments around the world.
• 04:41 And you start looking at this going, well, I can do this, but
• 04:44 it's getting really, really complicated.
• 04:46 And the file's getting really, really big, and I'm starting to get really,
• 04:50 really worried about maintaining this and the stability.
• 04:52 And that's what happens to some of these projects.
• 04:55 So sometimes you need to sit down and say hang on a second.
• 04:59 Do we really need to have all of this in one place or
• 05:03 can we actually break it down into individual components.
• 05:07 And I would say you know what?
• 05:08 One file can't do it all, one file can do a lot but don't try and
• 05:12 go overboard with it.
• 05:13 And the reason being is because Power BI actually gives us some flexibility,
• 05:17 we can build individual projects.
• 05:19 And when we get into the stage of building dashboards you'll see that you can
• 05:22 actually consolidate visuals from all kinds of different projects
• 05:26 into one central location.
• 05:28 So don't do it all in one place.
• 05:30 Be aware that you can split things up and
• 05:32 pull them back together again in a summary format later.
• 05:36 And that is a really important concept that will make more sense,
• 05:38 once you start seeing the dashboards build up.

Lesson notes are only available for subscribers.