How I created using Flow, Azure SQL DB, Cognitive Services & PowerBi

Last weekend I was thinking about how to save the tweets for PowerShell Conference Europe. This annual event occurs in Hanover and this year it is on April 17-20, 2018. The agenda has just been released and you can find it on the website

I ended up creating an interactive PowerBi report to which my good friend and Data Platform MVP Paul Andrew b | t added a bit of magic and I published it. The magnificent Tobias Weltner b | t who organises PSConfEU pointed the domain name at the link. It looks like this.

During the monthly #PSTweetChat

I mentioned that I need to blog about how I created it and Jeff replied

so here it is! Looking forward to seeing the comparison between the PowerShell and Devops Summit and the PowerShell Conference Europe 🙂

This is an overview of how it works


You will find all of the resources and the scripts to do all of the below in the GitHub repo. So clone it and navigate to the filepath

Create Database

First lets create a database. Connect to your Azure subscription

01 - subscription.png

Then set some variables

They should all make sense, take note that you need to set and uncomment the login and password and choose which IPs to allow through the firewall

Create a Resource Group

02 - resource group.png

Create a SQL Server

03 - create server.png

Create a firewall rule, I just use my own IP and add the allow azure IPs

03a - firewall rule.png

Create a database

04 - create database.png

I have used the dbatools module to run the scripts to create the database. You can get it using

Run the scripts

05 - Create Table Sproc.png

This will have created the following in Azure, you can see it in the portal

07 - portal.png

You can connect to the database in SSMS and you will see

06 - show table.png

Create Cognitive Services

Now you can create the Text Analysis Cognitive Services API

First login (if you need to) and set some variables

Then create the API and get the key

You will need to accept the prompt

08 -cognitive service

Copy the Endpoint URL as you will need it.Then save one of  the keys for the next step!

09 cognitiveservice key


Create the Flow

I have exported the Flow to a zip file and also the json for a PowerApp (no details about that in this post). Both are available in the Github repo. I have submitted a template but it is not available yet.

Navigate to and sign in

Creating Connections

You will need to set up your connections. Click New Connection and search for Text

16 - import step 3.png

Click Add and fill in the Account Key and the Site URL from the steps above

17 import step 5.png

click new connection and search for SQL Server

18 - import step 6.png

Enter the SQL Server Name (value of $AzureSQLServer) , Database Name , User Name and Password from the steps above

19 - import step 7.png

Click new Connection and search for Twitter and create a connection (the authorisation pop-up may be hidden behind other windows!)

Import the Flow

If you have a premium account you can import the flow, click Import

11 - import flow.png

12 - choose import.png

and choose the from the Github Repo

13 import step 1.png


Click on Create as new and choose a name

14 - import step 2.png

Click select during import next to Sentiment and choose the Sentiment connection

15 impot step 3.png

Select during import for the SQL Server Connection and choose the SQL Server Connection and do the same for the Twitter Connection

20 - import stpe 8.png

Then click import

21 - imported.png

Create the flow without import

If you do not have a premium account you can still create the flow using these steps. I have created a template but it is not available at the moment. Create the connections as above and then click Create from blank.

22 - importblank.png


Choose the trigger When a New Tweet is posted and add a search term. You may need to choose the connection to twitter by clicking the three dots

23 - importblank 1.png

Click Add an action

24 - add action.png

search for detect and choose the Text Analytics Detect Sentiment

25 - choose sentuiment.png

Enter the name for the connection, the account key and the URL from the creation of the API above. If you forgot to copy them

26 - enter details.png

Click in the text box and choose Tweet Text

27 - choose tweet text.png

Click New Step and add an action. Search for SQL Server and choose SQL Server – Execute Stored Procedure

28 - choose sql server execute stored procedure.png

Choose the stored procedure [dbo].[InsertTweet]

29 - choose stored procedure.png

Fill in as follows

  • __PowerAppsID__         0
  • Date                                 Created At
  • Sentiment                      Score
  • Tweet                              Tweet Text
  • UserLocation                 Location
  • UserName                      Tweeted By

as shown below

30 stored procedure info.png

Give the flow a name at the top and click save flow

31 flow created.png

Connect PowerBi

Open the PSConfEU Twitter Analysis Direct.pbix from the GitHub repo in PowerBi Desktop. Click the arrow next to Edit Queries and then change data source settings

32 change data source.png

Click Change source and enter the server (value of $AzureSQLServer) and the database name. It will alert you to apply changes

33 apply changes.png

It will then pop-up with a prompt for the credentials. Choose Database and enter your credentials and click connect

34 - creds.png

and your PowerBi will be populated from the Azure SQL Database 🙂 This will fail if there are no records in the table because your flow hasn’t run yet. If it does just wait until you see some tweets and then click apply changes again.

You will probably want to alter the pictures and links etc and then yo can publish the report

Happy Twitter Analysis

Dont forget to keep an eye on your flow runs to make sure they have succeeded.

One thought on “How I created using Flow, Azure SQL DB, Cognitive Services & PowerBi

Please feel free to comment on this post. All comments are moderated first before appearing on the site