Analyze in excel for power bi

Microsoft announced another awesome Power BI feature recently.  This Power BI Analyze in Excel feature was requested by Avi Singh from PowerPivotPro.com and was heavily supported as a requirement by the community. You can read more about how you can influence future releases at the bottom of this post.  But let’s get straight into this great new feature.

Power BI Analyze in Excel is available for the free service offering as well as the paid service. Note however if you want to upload a workbook and share it with someone else in your organisation so they too can Analyze in Excel, you will need the paid service.  But don’t despair- Power BI Pro is cheap and well worth the money.

Power BI Service

The Power BI Service is a modern version of SQL Server Analysis Services fully hosted in the Cloud by Microsoft.  What that means to you is you can have a powerful server to host your Power BI workbooks and distribute your reports without having to spend hundreds of thousands of dollars getting it all set up.  You can get started today for free with just a few clicks. https://powerbi.microsoft.com

Once you have your Power BI workbook loaded in the Cloud, you can even share a fully interactive version of the workbook publicly using the Publish to Web feature (demo’d below).  Note you should only do this with non-confidential data as this approach will make your data available publicly over the Internet.

You can view the Power BI report published on web here.

Once you have your data loaded up in Power BI, the world somehow just seems better.  Sometimes however you just want to get into the data and analyse it using good old fashioned Excel Pivot Tables.  Now can do that with the new Power BI Analyze in Excel feature.  To do this, you need to do the following:

• Log in to the Power BI Service with your account
• Navigate to the report in question in the left hand panel of the Power BI Service.
• Click on the ellipsis next to the report name (shown as 1 below)
• Click Analyze in Excel (2 below).

Updated Data Connector

The first time you do this, you will be prompted to download some updated software that is required for this feature to work.  You can prompt this manually from the drop down menu in the top right hand corner of PowerBI.com as shown below.

Make sure you select the correct version for your version of Excel.

• 32 bit version of Excel = Install x86 shown as 1 below
• 64 bit version of Excel shown as 2 below.

Once you have installed this, you can click the “Don’t show this again” box (shown as 3 above) to prevent this dialogue appearing each time you go through this process.

One thing to be aware of. After you click “Don’t show this again”, you can get to the download box again by clicking the downloads button (shown as 1 below) and then selecting “Analyze in Excel updates” (2 below).  However when you do this, it currently ONLY DOWNLOADS the 32 bit version regardless if you have 64 bit or not.  I have logged a bug with Microsoft and will remove this comment from this blog once the problem is fixed.

Building a Pivot Table

After you click the Power BI “Analyze in Excel” button, a small ODC file will be downloaded to your PC.  Keep an eye out for the download in your browser – mine is shown below.

When I click on the above ODC file, I get warning message as shown below.  Just click Enable.

After you click “Enable”, you will get a new blank Pivot Table connected to the Power BI Service as shown below.

Now you have a new blank workbook connected to your Power BI data model in the cloud.  You can build out a Pivot Table to analyse your data like any other Pivot Table, with a few minor differences.

If you take a close look at the Pivot Table Fields list on the right, you will see that there are now 2 types of field groups.  The first type (shown as 1 below) are the measures stored in your tables in Power BI.  These “measure tables” are indicated by the Sigma symbol.  The second type (shown as 2 below) are the actual tables (and columns) from the data model.  You select the Values for your Pivot Table from the measures tables (1) and the Rows/Columns/Filters/Slicers from the actual tables (2).

Now you just build out your Pivot Table as normal as shown below.

Thin Workbooks

When you save this file, you will notice something really special.  This is a “Thin Workbook”, in my example it is just 24kb (shown below).

The data does not live in the workbook – it is only in the cloud.  The only data that is in the workbook is the data that is visualised inside the Pivot Table(s).  But there is a cost to this approach – you must have a live connection to the Internet to be able to interact with the data in the Pivot Table.  Personally I think this is a small price to pay for all of the benefits, however if you want to interact with your workbook in a location that doesn’t have an Internet connection, then this is something you will need to be aware of.

Also note that now the data is in the cloud there will be a small amount of latency (delay) when clicking on the Pivot Table compared to a local workbook on your PC.  This will be most noticeable for small workbooks that are almost instantaneous on your PC.  These small fast workbooks will move from being “instantaneous” on your PC to being “sub second” in the cloud – which is not that bad in my view.  If you have very large workbooks on your PC (say 300MB +), or if you have 32 bit Excel on your PC, then you may actually notice an overall improvement with these thin workbooks.  This is because the processing of the cube is now pushed to the Power BI Service which has lots of powerful processors to complete the task.

Distributing Thin Workbooks

Now that you have this “Thin” Excel workbook, you can distribute the Thin Workbook to other users and they can also interact with the data live directly in the Excel workbook connected to the cloud.  You don’t need to distribute the ODC file, just the updated and saved Thin Workbook.  There are some pre-requisites for this to work, including:

1. You will need to send them the Thin Workbook once the connection has been established (of course).
2. They will need their own Power BI Service Account.
3. They will need access to the shared data via the Power BI Service, so you will need to share the Power BI report with them too from within Power BI.
4. They will need to install the updated “Analyze with Excel update” described earlier.

I have tested distributing this Thin Workbook to a Mac user with Excel 2011 and unfortunately it didn’t work. I haven’t tested it with Excel 2016 for Mac but would like to hear from anyone who does.

Better Than Export to Excel

Rob Collie has a favourite joke.  Qn. What is the 3rd most common button in all BI tools?  Ans. Export to Excel (3rd after 1. OK and 2. Cancel).  But Power BI Analyze with Excel is so much more than “Export to Excel”.  The reason of course is that you don’t have to take a copy of your data to work with it.  You create a live link to the data source and hence when the data is updated in the future, your “Analyze with Excel” workbook will also update.

Cube Formulas Work Too

If you like using Cube Formulas, the good news is they will work too.  When you type a cube formula (like shown below), you will be given a choice of data connections (1 being the Power BI service and 2 being the data model in the current workbook).

So that got me wondering… Can I have 2 data models for a workbook, 1 in the Cloud and a second in the workbook?  The answer is yes.  I have 2 cube formulas below, the first is coming from the data model in the current workbook and the second is coming from the Power BI service connected to the same workbook.

Common Errors

New Versions of Connector

The connector is being updated all the time.  If you see this error (or any other connection error for that matter)

The connection failed because user credentials are needed and Sign-In UI is not allowed

The first thing you should do is make sure you have the latest version of the connector installed.  You can find the latest update online at the Power BI Service as shown below.

Forbidden Error

If you have multiple Power BI accounts, you may come across the following error “The HTTP server returned the following error: Forbidden.”  In fact some users that don’t have multiple accounts have also experienced this error.

The issue is caused by “swapping” between accounts and potentially the login process trying to log into the wrong account.  There is a manual work around to fix this problem.

1. Navigate to the ODC file you downloaded.
2. Right click the ODC file and edit it in Notepad.
3. Find the section that starts with <odc:ConnectionString>
4. Add the following text immediately after this string

   User ID =

   [email protected];
5. Save the file, and then double click to open it again.

You should now be taken to the correct login screen to give you access to the data. I had this very issue myself when I first tried to use this feature.  I logged a thread at community.powerbi.com and got immediate help from the Microsoft Power BI support team to help me resolve the issue.  If you are not a member of the Power BI community, you are missing out – sign up and get involved.

Hopefully a more elegant solution will be delivered in the future – I’m sure it will.

Initialization of the data source failed

You may  see this error, particularly the first time you try to do this.

“Initialization of the data source failed. Check the database server or contact your database administrator. Make sure the external database is available, and then try the operation again. If you see this message again, create a new data source to connect to the database.”

This is normally related to you not having the correct data connector installed for your version of Excel.  Make sure you have installed the correct (64 bit or 32 bit) data connector as covered earlier in this post.

How to Influence Future Power BI Updates

Power BI is rapidly improving each week as the crack Microsoft Power BI team focuses on delivering the new features that its customers most want.  If you want to influence what features get built out next, then do what Avi did and create an idea at http://ideas.powerbi.com/ .  Or you can simply look at other people’s ideas and vote for the ones that you want most.

If you want to help me with one of my passions, you can vote for this idea here.

https://ideas.powerbi.com/forums/265200-power-bi-ideas/suggestions/12464280-individual-custom-links-for-pinned-live-pages

What I think is really needed is a way to build a proper “web page like” menu structure so you can build out Power BI with a “web browsing experience”.  The current approach of pinning tiles is OK in limited scenarios, but is not really “complete” enough to deliver a comprehensive reporting tool.  I think we need a proper menu/navigation editing solution like in SharePoint where the curator can build out a proper menu structure to deliver the reports to users in a custom and meaningful way.

Power BI and Excel are longtime friends of each other, not only because Power BI components come from add-ins introduced in Excel but also because of how these two tools interact with each other from the Power BI Service. This post is not about using Power Query or Power Pivot components in Excel. This article and video discuss the interaction between Excel and Power BI through the service.

Power BI and Excel integration through the service allows the user to use Excel as their slicing and dicing tool while connected to a live Power BI dataset. On the other hand, you can pin a range of cells from an Excel document into the Power BI dashboard. Excel files also can be uploaded to the Workbook tab of the Power BI service. In this post, you will learn the below ways that Excel and Power BI interact with each other through Power BI Service;

• Analyze in Excel
• Publish to Power BI from Excel
• Import Excel into Power BI Desktop

Video

Analyze in Excel

Every company will find some users with very good experience and skillset in Excel. Excel users can still use Excel to connect to the Power BI dataset and use Excel features such as PivotTable and PivotChart to slice and dice the data. The connection to the Power BI dataset would be a live connection, which means whenever users refresh the Excel file, they will get the most up-to-date data from the Power BI service. Now let’s see how Analyze in Excel works in action.

Analyze in Excel from Power BI Service

You can initiate the Analyze in Excel from a Power BI Service report or a dataset. Log in to the Power BI service and click on the more options of a dataset (or a report). Click on the Analyze in Excel.

Power BI will generate an Excel file in OneDrive and create a connection from that Excel file to the Power BI dataset. You will get a notification when the Excel file is ready. (If you don’t have OneDrive for Business in your tenant, then clicking on Analyze in Excel will download the workbook to your local computer)

The opening of this file, by default, is in Excel online (web version). However, you can download the file and store it locally. Once you open the file, you will get a warning about the data coming from outside of this workbook (Power BI Dataset)

Once you click Yes, you will see an Excel workbook with a PivotTable, which is the data sourced from the Power BI dataset in the service. Analyze in Excel will use the same Power BI account username and password that you had access to the report from it (because the Power BI account is an Office 365 account).

Drag data fields into the slicing and dicing area (right under the fields pane), and you will see a result in PivotTable. This result is fetched live from the Power BI dataset in the Power BI service.

Implicit Measures won’t work in Excel.

Implicit measures are a measure that Power BI creates automatically. Power BI automatically applies auto summarization on numeric fields (that haven’t been part of a relationship). Behind the scenes, Power BI creates a measure for those fields; these measures are called Implicit measures. These are a measure that you can see a small Sum or Sigma icon beside their name in the Power BI Desktop.

The implicit measure cannot be used in the Analyze in Excel feature; if you try dragging them in PivotTable, you will see individual values instead of the aggregation or measure result.

However, if you create explicit measures (which are DAX measures created by you), similar to the screenshot below;

Then you can use them in PivotTable as a normal measure and see the correct result.

If you consider some users to use Excel as their front-end tool to connect to Power BI Models, You have to consider creating the Explicit measure.

Excel is Connected Live to the Power BI Model in the Service

The wonderful thing about the Excel connection to the Power BI Service is that the connection is Live. Live connection means Excel fetches the data directly from the dataset in the Power BI service. Anytime you refresh the Excel file, you get the most up-to-date data from the service. This feature is completely different from Export to Excel. The export to Excel option you see on visuals in the Power BI service only downloads data offline. However, Analyze in Excel is an online and live connection to the dataset.

You can check the connection properties in the Data tab, under the Connections, Properties section.

The Connection properties will have an Azure address in it with the ID of the dataset in the Power BI service. You can even use this connection in any other excel file to connect to the same dataset.

Get Data from Power BI Dataset

Another very similar approach to Analyzing in Excel is getting data from the Power BI dataset in Excel. If you have an Excel file and you want to do some data analysis on the data of a Power BI dataset, you can go to the Data tab, under getting Data, Under Power Platform, Choose Dataset.

This will show you the list of Power BI datasets that you have access to, and you can select to get data from any of those.

This will create a live connection to the Power BI dataset (Similar to a thin report).

Get Data from Power BI Dataflow

Using Excel, you can also connect to a Power Platform Dataflow and get data from it.

The Navigator window will show you both Power Platform dataflows (under Environments) and Power BI dataflow (under workspaces);

Why Analyze in Excel is Better than Export Data?

There are many reasons why Analyzing in Excel is a better option than exporting data. Let’s explore some of those here.

Analyze in Excel is Live Connection, Export Data is Offline, outdated

As soon as you export data from a Power BI report or dataset, you create a snapshot of the data then. If the report refreshes the next day, your exported data is no longer up-to-date. It is offline and outdated.

However, when you use Analyze in Excel. Excel’s data is fetched live from the Power BI dataset in the service. You can see that in the connection properties of Exel. You can check the connection properties in the Data tab, under the Connections, Properties section. The Connection properties will have an Azure address in it with the ID of the dataset in the Power BI service. You can even use this connection in any other excel file to connect to the same dataset.

Export Data is limited to 150,000 Rows; analyze in Excel Not!

Export data has a limitation on the number of rows. Analyze in Excel doesn’t. You are connected to the model live. It looks like a live connection to SSAS tabular; you can do whatever you want with the data.

The export is limited to 150K rows if you export to Excel and 30K rows for CSV files.

Export Data is limited to specific visuals and fields used in that visual

When you use Export Data, you only do it from a specific visual. And fields in that visual (or related fields to that) are exported.

When you use analyze in Excel, you will have access to all the tables, columns, and their calculated fields and measures in Excel:

Analyze in Excel is the Entire model: all tables and calculations

Export Data is not Secure. Analyze in Excel uses Power BI account credentials.

You might export data using an account that has access to everything. And then share the exported data file with someone who should NOT have access to the data. Because the security and the data are decoupled when you use the export data, then there is no security around it. You have to be careful to secure the exported data yourself.

However, when you use Analyze in Excel, You have to log in with your Power BI account and see the data only if you have access to it.

Analyze in Excel supports Row-Level Security.

Because users log in from Excel, when they use the Analyze in Excel feature will see only part of the data they are allowed to see. That means if there is row-level security enabled on the data. Users will see only their part of the data, nothing more. (as long as they don’t have Edit access on the dataset)

Analyze in Excel uses the concept of Shared Datasets

When you use the Analyze in Excel feature, you use the shared Power BI dataset from the Excel front end. Whenever the dataset refreshes, you have the new data. You will get the updates if the dataset gets a new table or field.

As the above screenshot shows, the Analyze in Excel works well with multi-layered architecture for Power BI development. The visualization layer can be Power BI reports, paginated reports, or even Excel.

Publish to Power BI from Excel

Not only can Excel read data from the Power BI Service, but also it can publish data into Power BI. This is possible in Excel 2016 editions or Office 365 editions of Excel. Once you click the Publish button on the File menu, you will see an option to Publish into Power BI.

Upload your workbook to Power BI

If you have an Excel workbook file, which you want to keep beside your Power BI objects (reports, dashboards, and datasets), then this option is for you. You can upload the workbook into the Power BI workspace using this option. Once the process is completed, The excel workbook will be accessible from the Power BI workspace and can be opened and edited in Excel online.

The Excel workbook, as mentioned, can be edited or viewed online.

Export workbook data to Power BI

Here comes the exciting feature; not only can you upload the excel file as a workbook, but also, you can export it as a Power BI dataset. To do this, you must have data in the Excel file as tables. The tables’ data will then be exported into a Power BI dataset from Excel (and it can be updated by updating the Excel file).

After these publish options, you will see a short message showing the publish process to the Power BI service;

These two options might not be very commonly used, but they certainly will bring an interesting aspect of integration between Excel and Power BI.

Import Excel into Power BI Desktop

We cannot discuss all Excel and Power BI integration options without mentioning the most important developer option. You can import an Excel model which includes a Power Pivot model into Power BI. The Import into Power BI doesn’t mean Get Data from Excel, it means the entire model, including all the tables, relationships, and calculations, will be imported into a Power BI report. The scenario above is a very handy option when you have an existing Power Pivot model in Excel.

Importing an Excel workbook has a very easy and straightforward process. The only important note in this process is that your Excel file should not be password protected.

Summary

In this article, you’ve learned methods by that Excel, and Power BI work together. Each of the modes of integration will give you some features. Analyzing in Excel will allow you to slice and dice the data model of the Power BI service easily from Excel through PivotTable and PivotChart. Excel can publish the entire workbook or the data part as a dataset into Power BI. Both of the features mentioned above will be highly well received by your business users who are good with Excel as a front-end tool. Last but not least, you can import an entire Excel Power Pivot model with all tables, relationships, and calculations into a Power BI report instead of re-creating it.