Monthly Archives: April 2015

Part 2: Real time dashboard using a Raspberry PI and Power BI

Part 2: Using the Power BI API to make a real time dashboard

So you have set up your Azure Active Directory and created a Power BI provision for your organization? If not, you can have a look at part one of this series.

Before we start playing with the Raspberry PI let’s create a simple console application to see that we have set up things correctly in Azure Active Directory in the last post and are able to have real time updates on a dashboard in Power BI. Nothing advanced, but it is a nice test to see if we have set everything up correctly with our Azure account and Power BI. What we will do is create a dataset if it does not already exist in Power BI and insert rows of data into that dataset which will then update the dashboard in real time.

Creating a Power BI application

If you want to take the easy way out to see if you have set up everything correctly you can download a test code made by the MSDN team on github here: https://github.com/PowerBI/getting-started-for-dotnet. After downloading this code you need to insert your own Client ID that you get from your application page in Azure. You also need to set the redirect URI to the same as your application, if you don’t do this you will get an error message like the one below.

BadRequest

When you run the code you will see a green star-thingy appear to the right of your newly created dataset. After the dataset is created you can go and explore that dataset, even if it’s empty, and create a report. From the report you then pin report elements to a dashboard and when you run your application you will see the elements update real time. Nice!

NewDataset

I would suggest that you, either if you choose the easy way out or make a new application  from scratch, do some playing around and create your own dataset and insert some rows into it so you get a better understanding on what is happening and how the Power BI API works. Getting a good understanding of how datasets and inserting rows work will be useful later. We will not be able to reuse the dataset in the example code when using our Raspberry PI so we need to be able to create our own datasets, turn them into JSON and then insert some data. What you create is completely up to you and your imagination. I made a new application from scratch with a simple timer that  every second writes how many seconds it has been running, along with a timestamp, to Power BI. Does it provide great insight? Do I change how we look at this world? No, but it has provided me with some valuable hands on experience with the Power BI API. Other examples I’ve seen people doing is for example looking at clicks on a website and what device the clicks come from. Again it is completely up to you what you want to do.

Next up we will start to play with the Raspberry PI and then connect it to Power BI.

Dashboard

NOTE: When writing to a Power BI dataset there is, at the point when this post was written, no option to export a dataset from the Power BI Preview. This means that if you want to be able to use the data in another format than Power BI you need to store it another place as well. I am going to create an Azure SQL Database to store my data when using the Raspberry PI in case I want to use it other places other than PowerBI. Maybe you should consider doing the same?

Other parts of the series can be found here:

Part 1: Real time dashboard using a Raspberry PI and Power BI

Why I really like Datazen as a visualization tool

I was really excited when I saw that Microsoft had bought Datazen. I have been able to use Datazen for a while already, and I really like the product. At the Gurus of Business Intelligence conference in Oslo in 2014 I was lucky enough to be able to have a session together with Marc Reguera, @marcreguera, where I spoke about mobile and touch based reporting using Datazen as the example app.

Now why do I like Datazen? First of all I think they have a clear vision, a vision I like. They want to create great data visualizations for mobile and touch based devices which are easily designed and shared with others. It is not a data discovery tool like Qlik or Tableau so you need to have your data organized in order to get the full effect. But when you are ready to create visualizations and share those with your coworkers it is really good in my opinion. You have predefined graphical elements you drag and drop onto your design surface and connect your data to. You also have the possibility to create custom maps and color palettes with backgrounds to get the right feel for your dashboards.

Datazen has three major components. You need a Datazen server in order to get data from other sources than local Excel files. When you have created a dashboard you are also able to deploy your dashboard to the Datazen server so it can be consumed by others. The server also grants you access to a control panel where you can setup KPIs, set permissions, set up connections to your data sources and create data views which can then be imported into your dashboards.

This brings us to the other two parts of the Datazen stack, the viewer and publisher apps. The publisher app is only available on Windows 8 devices at this point. This is where you import your data and create your dashboards before you can publish them to the server. The publisher itself is also made for touch so using f.ex a Surface to create visualizations on the go, or in a demo environment, is a real joy! It is quite fast to setup a dashboard design and connect your data to the graphical elements you have available. In my eyes this is a really important feature because it allows you to quickly iterate over different designs. I don’t know how many times a client has come back and said “But that’s not how I visualized the report in my head”. Now you can sit with your client and quickly make a new version of the same dashboard right there and then. I have several times prepared a dashboard to present in a demo, but have found myself creating a dashboard on the go instead. Because it is easy to do so. Because I can. You can even hand over your computer to the client and let them create the dashboard during your demo!

The viewer app lets you consume dashboards deployed to a Datazen server you have access and is available for all mobile platforms as a native app. If you are used to Android you will still have the Android feel unlike some apps that have been made to fit all platforms, but rarely do. When you design a dashboard you create several versions of the same dashboard to fit different resolutions. A dashboard made for a phone may have to emphasize different things than one made for a laptop, simply because some graphical elements, like a map, is perhaps not that suited for a small screen. Datazen also gives you offline access to your dashboard which I think is absolutely essential for all apps.

You can download the Datazen publisher for free from the Windows 8 store and use local Excel files to create your own stunning dashboards right away. If you want to share those dashboards or get data from your databases or other sources, you will need access to a Datazen server. When the announcement was made Microsoft also announced that SQL Server Enterprise Edition customers with version 2008 or later and Software Assurance are entitled to download the Datazen Server software at no additional cost. Give it a go. It is a really well made product, and a lot of fun to play with! The road from raw data to presenting them on a well made dashboard on someone else’s mobile phone is getting shorter!

For more information go to http://www.datazen.com/.

Example of a dashboard made with Datazen

 

 

Part 1: Real time dashboard using a Raspberry PI and Power BI

Part 1: Setting up Azure Active Directory and registering your application

When the Power BI preview finally was released I was really intrigued by the Power BI API which among other things allow you to create dashboards that are updated in real time. Now, I am expecting my first child in under a month now and I thought that why not combine my background in cybernetics and embedded systems with my present, business intelligence. I have bought a Raspberry PI and I invite you to join my mission to hopefully create a baby sound monitor that will be updating a Power BI dashboard in real time to show how often, how loud and how long my baby is crying. Hopefully we will have a quiet baby that never screams and I’ll have to use this for registering baby laughter instead, but who knows right? You might not have a baby on its way, but maybe you have some other IoT project you want to do or just want to try out the Power BI API. Either way you’re invited to embark on this journey together with me!

I will make this into a series of post, but the journey starts with setting up your Azure Active Directory and getting to know the Power BI preview where a dashboard will get updated real time using the Power BI API later. In order to use the Power BI API you need to register the application you will be making in your Azure Active Directory. I also made a new user that I am going to use for my Power BI site.

Creating an Azure Active Directory

When you log into your Azure Management Portal you will find Active Directory on the navigation bar on the left. Click on Active Directory and the new button, choose App Services, Directory and click on Custom Create to create a new AD. Give it a name, a domain name and select the country for your directory. The country can’t be changed after the directory is made and affects which data center your AD will be located in so choose wisely, or just go with the country you’re in as I did.

NewAD
AddDirectory

When your directory is created it will show up under your Azure Active Directories. I have made a new user that I am going to use with Power BI and then signed up for a preview over at www.powerBI.com using this user from my Azure Active Directory. It might take some time to get a Power BI provision up for your Azure Active Directory  organization so I would do this step before you read the rest of this post.

Registering your Power BI app

Before you can create your Power BI application you need to register your app in your Azure Active Directory. Click on the directory you created, go to applications and press add. You want to “add an application my organization is currently developing” and you want it to be a “native client application”. You are asked for a redirect URI where you can use https://login.live.com/oauth20_desktop.srf. Make sure when you start writing your program to use the same redirect URI in your code.

When your application is up you need to set some permissions for it. Click on your application, go to configure, scroll to the bottom of the page and press Add Application. Select “Power BI Service” and close the window. Power BI Service will show up under “Permissions to other applications”, but before you are done you have to check all three checkboxes in the dropdown “Delegated Permissions”.

PowerBIPermissions
PowerBIPermissions2

And with that step: Congratulations! You are now ready to copy/paste the Client ID from your application and start using it together with the Power BI API. We will continue on that later. Feel free to play with the Power BI preview as well to familiarize yourself by f.ex uploading an Excel file as a data source and make some reports and dashboards. Personally I am looking forward to having more visualizations as I almost felt the word preview should have been emphasized more in this Power BI preview.

PowerBIProperties
Other parts of the series can be found here:

Part 2: Real time dashboard using a Raspberry PI and Power BI