Data Setup (Bring Your Own Tenant)

Prerequisites

• An Azure subscription where you can create resources (Owner or Contributor).
• A Microsoft Fabric capacity or an active Fabric trial. Start one at app.fabric.microsoft.com → Account manager → Start trial.
• The seed-data scripts from the lab repo: github.com/amruthasatishkumar/isv-summit-data-lab (the seed-data/ folder).
• Local tooling: Azure CLI, Python 3.10+, and sqlcmd (or the Azure portal Query editor).

What you are creating

Five Azure resources in one resource group, plus a Fabric workspace. Each maps to a part of the lab:

Step 1 · Create a resource group

Sign in and create one resource group to hold everything. Pick a region close to you and use it consistently.

az login

# pick your own names and region
$rg     = "rg-urbanpulse-lab"
$region = "eastus2"

az group create --name $rg --location $region

Step 2 · Azure SQL Database + seed

Create a logical SQL server and a database, open the firewall to your client, then run the bundled setup script that creates the schema, seeds the data, and enables Change Tracking for mirroring.

1. Create the server and database:

   $server = "urbanpulse-$(Get-Random)"   # must be globally unique
   $db     = "UrbanPulse"
   $sqlUser = "labadmin"
   $sqlPass = ""
   
   az sql server create --name $server --resource-group $rg --location $region `
       --admin-user $sqlUser --admin-password $sqlPass
   
   az sql db create --resource-group $rg --server $server --name $db `
       --service-objective S0
   
   # allow your client IP through the firewall
   $ip = (Invoke-RestMethod https://api.ipify.org)
   az sql server firewall-rule create --resource-group $rg --server $server `
       --name AllowMyClient --start-ip-address $ip --end-ip-address $ip

2. Download the setup script and run it against your database. It is idempotent, so it is safe to re-run.

   sqlcmd -S "$server.database.windows.net" -d $db -U $sqlUser -P $sqlPass `
       -i urbanpulse-sql-setup.sql

Step 3 · Azure Cosmos DB + seed

Create a Cosmos DB (NoSQL) account, then load the regions and trainRoutes documents with the bundled Python loader. Download the bundle below - it contains the loader plus the two JSON data files. The loader creates the database and both containers if they are missing and upserts every document, so it is safe to re-run.

1. Create the account:

   $cosmos = "urbanpulse-cosmos-$(Get-Random)"   # globally unique
   
   az cosmosdb create --name $cosmos --resource-group $rg `
       --locations regionName=$region --kind GlobalDocumentDB

2. Grab the endpoint and primary key:

   $cosmosEndpoint = az cosmosdb show --name $cosmos --resource-group $rg --query documentEndpoint -o tsv
   $cosmosKey      = az cosmosdb keys list --name $cosmos --resource-group $rg --query primaryMasterKey -o tsv

3. Extract the bundle, then run the loader from the extracted folder (it reads regions.json and trainRoutes.json from next to load_cosmos.py):

   pip install azure-cosmos
   
   $env:COSMOS_ENDPOINT = $cosmosEndpoint
   $env:COSMOS_KEY      = $cosmosKey
   $env:COSMOS_DATABASE = "urbanpulse"
   
   python load_cosmos.py

Step 4 · Azure Storage + facility catalog

Create a storage account and a public blob container named facilitycatalog (that exact name is referenced in M2), then upload the facility catalog Parquet file and the five facility photos.

1. Create the account and container:

   $storage = "urbanpulse$(Get-Random)"   # 3-24 lowercase chars, globally unique
   
   az storage account create --name $storage --resource-group $rg `
       --location $region --sku Standard_LRS
   
   az storage container create --account-name $storage `
       --name facilitycatalog --public-access blob

2. Download the data bundle and extract it. You should have facility_catalog.parquet next to a facilities/ folder of five JPGs.
3. Upload both into the container:

   az storage blob upload-batch `
       --account-name $storage `
       --destination facilitycatalog `
       --source .\urbanpulse-lab-data `
       --overwrite

Step 5 · Event Hubs + streaming simulators

Create an Event Hubs namespace and three hubs, then run the Python simulators to produce live events for M3. The hub names must match exactly - the M3 KQL ingestion mapping reads named JSON fields and silently drops anything that does not match.

1. Create the namespace and hubs:

   $ehNamespace = "urbanpulse-eh-$(Get-Random)"   # globally unique
   
   az eventhubs namespace create --name $ehNamespace --resource-group $rg `
       --location $region --sku Standard
   
   foreach ($hub in "medicalvitals","medicalmovement","metrotrain") {
       az eventhubs eventhub create --resource-group $rg `
           --namespace-name $ehNamespace --name $hub --partition-count 2
   }

2. Get a per-hub connection string (it must end with ;EntityPath=<hub>):

   $rule = "RootManageSharedAccessKey"
   foreach ($hub in "medicalvitals","medicalmovement","metrotrain") {
       $base = az eventhubs eventhub authorization-rule keys list `
           --resource-group $rg --namespace-name $ehNamespace `
           --eventhub-name $hub --name $rule --query primaryConnectionString -o tsv 2>$null
       Write-Host "$hub => $base"
   }

   If the per-hub rule does not exist yet, create one per hub with az eventhubs eventhub authorization-rule create ... --rights Send Listen, or use the namespace-level RootManageSharedAccessKey and append ;EntityPath=<hub> yourself.

3. Run the three simulators from seed-data/eventhub/ in separate terminals:

   pip install azure-eventhub
   
   $env:EVENTHUB_VITALS_CONN_STR   = "...;EntityPath=medicalvitals"
   python seed-data\eventhub\simulate_medicalvitals.py
   
   $env:EVENTHUB_MOVEMENT_CONN_STR = "...;EntityPath=medicalmovement"
   python seed-data\eventhub\simulate_medicalmovement.py
   
   $env:EVENTHUB_TRAIN_CONN_STR    = "...;EntityPath=metrotrain"
   python seed-data\eventhub\simulate_metrotrain.py

Step 6 · Create a Fabric workspace

1. Go to app.fabric.microsoft.com.
2. Select Workspaces → + New workspace, name it (for example UrbanPulse-Lab), and assign it to your Fabric capacity or trial.
3. Continue with Module 1 · Workspace Tour. When a module asks for a server name, key, or connection string, use the values from the resources you just created instead of the shared lab credentials.