cortana-analytics-playbook-vehicle-telemetry-powerbi.md

title	description	services	documentationcenter	author	manager	editor	ms.assetid	ms.service	ms.workload	ms.tgt_pltfrm	ms.devlang	ms.topic	ms.date	ms.author
Vehicle telemetry analytics solution template Power BI Dashboard setup instructions | Microsoft Docs	Use the capabilities of Cortana Intelligence to gain real-time and predictive insights on vehicle health and driving habits.	machine-learning		bradsev	jhubbard	cgronlun	aaeb29a5-4a13-4eab-bbf1-885690d86c56	machine-learning	data-services	na	na	article	12/16/2016	bradsev

Vehicle telemetry analytics solution template Power BI Dashboard setup instructions

This menu links to the chapters in this playbook.

[!INCLUDE cap-vehicle-telemetry-playbook-selector]

The Vehicle Telemetry Analytics solution showcases how car dealerships, automobile manufacturers and insurance companies can leverage the capabilities of Cortana Intelligence to gain real-time and predictive insights on vehicle health and driving habits to drive improvements in the area of customer experience, R&D and marketing campaigns. This document contains step by step instructions on how you can configure the Power BI reports and dashboard once the solution is deployed in your subscription.

Prerequisites

1. Deploy the Vehicle Telemetry Analytics solution by navigating to https://gallery.cortanaanalytics.com/SolutionTemplate/Vehicle-Telemetry-Analytics-3
2. Install Microsoft Power BI Desktop
3. An Azure subscription. If you don't have an Azure subscription, get started with Azure free subscription
4. Microsoft Power BI account

Cortana Intelligence Suite Components

As part of the Vehicle Telemetry Analytics solution template, the following Cortana Intelligence services are deployed in your subscription.

• Event Hubs for ingesting millions of vehicle telemetry events into Azure.
• Stream Analytics for gaining real-time insights on vehicle health and persists that data into long-term storage for richer batch analytics.
• Machine Learning for anomaly detection in real-time and batch processing to gain predictive insights.
• HDInsight is leveraged to transform data at scale
• Data Factory handles orchestration, scheduling, resource management and monitoring of the batch processing pipeline.

Power BI gives this solution a rich dashboard for real-time data and predictive analytics visualizations.

The solution uses two different data sources: Simulated vehicle signals and diagnostic dataset and vehicle catalog.

A vehicle telematics simulator is included as part of this solution. It emits diagnostic information and signals corresponding to the state of the vehicle and driving pattern at a given point in time.

The Vehicle Catalog is a reference dataset containing VIN to model mapping

Power BI Dashboard Preparation

Deployment

Once the deployment is completed, you should see the following diagram with all of these components marked in GREEN.

• To navigate to the corresponding services to validate whether all of these have deployed successfully, click the arrow on the upper right of the green nodes.
• To download the data simulator package, click the arrow on the upper right on the Vehicle Telematics Simulator node. Save and extract the files locally on your machine.

Now, you are ready to configure the Power BI dashboard with rich visualizations to gain real-time and predictive insights on vehicle health and driving habits. It takes about 45 minutes to an hour to create all the reports and configure the dashboard.

Setup Power BI Real-Time Dashboard

Generate simulated data

1. On your local machine, go to the folder where you extracted the Vehicle Telematics Simulator package.
2. Execute the application CarEventGenerator.exe.
3. It emits diagnostic information and signals corresponding to the state of the vehicle and driving pattern at a given point in time. This is published to an Azure Event Hub instance that is configured as part of your deployment.

Start the real-time dashboard application

The solution includes an application that generates a real-time dashboard in Power BI. This application listens to an Event Hub instance, from which Stream Analytics publishes the events continuously. For every event that this application receives, it processes the data using a Machine Learning Request-Response scoring endpoint. The resultant dataset is published to the Power BI push APIs for visualization.

To download the application:

1. Click the Power BI node on the diagram view and click the Download Real-time Dashboard Application’ link on the properties pane.

2. Extract and save the application locally

3. Execute the application RealtimeDashboardApp.exe

4. Provide valid Power BI credentials, sign in and click Accept

   

   

Configure Power BI reports

The real-time reports and the dashboard take about 30-45 minutes to complete. Browse to http://powerbi.com and login.

A new dataset is generated in Power BI. Click the ConnectedCarsRealtime dataset.

Save the blank report using Ctrl + s.

Provide report name Vehicle Telemetry Analytics Real-time - Reports.

Real-time reports

There are three real-time reports in this solution:

1. Vehicles in operation
2. Vehicles Requiring Maintenance
3. Vehicles Health Statistics

You can choose to configure all the three real-time reports or stop after any stage and proceed to the next section of configuring the batch reports. We recommend you to create all the three reports to visualize the full insights of the real-time path of the solution.

1. Vehicles in operation

Double-click Page 1 and rename it to “Vehicles in operation”

Select vin field from Fields and choose visualization type as “Card”.

Card visualization is created as shown in figure.

Click the blank area to add new visualization.

Select City and vin from fields. Change visualization to “Map”. Drag vin in values area. Drag city from fields to Legend area.

Select format section from Visualizations, click Title and change the Text to “Vehicles in operation by city”.

Final visualization looks as shown in figure.

Click the blank area to add new visualization.

Select City and vin, change visualization type to Clustered Column Chart. Ensure City field in Axis area and vin in Value area

Sort chart by “Count of vin”

Change chart Title to “Vehicles in operation by city”

Click the Format section, then select Data Colors, Click the “On” to Show All

Change the color of individual city by clicking on color icon.

Click the blank area to add new visualization.

Select Clustered Column Chart visualization from visualizations, drag city field in Axis area, Model in Legend area and vin in Value area.

Rearrange all visualization on this page as shown in figure.

You have successfully configured the “Vehicles in operation” real-time report. You can proceed to create the next real-time report or stop here and configure the dashboard.

2. Vehicles Requiring Maintenance

Click to add a new report, rename it to “Vehicles Requiring Maintenance”

Select vin field and change visualization type to Card.

We have a field named “MaintenanceLabel” In the dataset. This field can have a value of “0” or “1”.” It is set by the Azure Machine Learning model provisioned as part of solution and integrated with the real-time path. The value “1” indicates a vehicle requires maintenance.

To add a Page Level filter for showing vehicles data, which are requiring maintenance:

1. Drag the “MaintenanceLabel” field into Page Level Filters.
   
2. Click Basic Filtering menu present at bottom of MaintenanceLabel Page Level Filter.
   
3. Set its filter value to “1”
   

Click the blank area to add new visualization.

Select Clustered Column Chart from visualizations

Drag field Model into Axis area, Vin to Value area. Then sort visualization by Count of vin. Change chart Title to “Vehicles requiring maintenance by model”

Drag vin fields into Color Saturation present at Fields section of Visualization tab

Change Data Colors in visualizations from Format section
Change Minimum color to: F2C812
Change Maximum color to: FF6300

Click the blank area to add new visualization.

Select Clustered column chart from visualizations, drag vin field into Value area, drag City field into Axis area. Sort chart by “Count of vin”. Change chart Title to “Vehicles requiring maintenance by city”

Click the blank area to add new visualization.

Select Multi-Row Card visualization from visualizations, drag Model and vin into the Fields area.

Rearranging all of the visualization, the final report looks as follows:

You have successfully configured the “Vehicles Requiring Maintenance” real-time report. You can proceed to create the next real-time report or stop here and configure the dashboard.

3. Vehicles Health Statistics

Click to add new report, rename it to “Vehicles Health Statistics”

Select Gauge visualization from visualizations, then drag the Speed field into Value, Minimum Value, Maximum Value areas.

Change the default aggregation of speed in Value area to Average

Change the default aggregation of speed in Minimum area to Minimum

Change the default aggregation of speed in Maximum area to Maximum

Rename the Gauge Title to “Average speed”

Click the blank area to add new visualization.

Similarly add a Gauge for average engine oil, average fuel, and average engine temperate.

Change the default aggregation of fields in each gauge as per above steps in “Average speed” gauge.

Click the blank area to add new visualization.

Select Line and Clustered Column Chart from visualizations, then drag City field into Shared Axis, drag speed, tirepressure and engineoil fields into Column Values area, change their aggregation type to Average.

Drag the engineTemperature field into Line Values area, change the aggregation type to Average.

Change the chart Title to “Average speed, tire pressure, engine oil and engine temperature”.

Click the blank area to add new visualization.

Select Treemap visualization from visualizations, drag the Model field into the Group area, and drag the field MaintenanceProbability into the Values area.

Change the chart Title to “Vehicle models requiring maintenance”.

Click the blank area to add new visualization.

Select 100% Stacked Bar Chart from visualization, drag the city field into the Axis area, and drag the MaintenanceProbability, RecallProbability fields into the Value area.

Click Format, select Data Colors, and set the MaintenanceProbability color to the value “F2C80F”.

Change the Title of the chart to “Probability of Vehicle Maintenance & Recall by City”.

Click the blank area to add new visualization.

Select Area Chart from visualization from visualizations, drag the Model field into the Axis area, and drag the engineOil, tirepressure, speed and MaintenanceProbability fields into the Values area. Change their aggregation type to “Average”.

Change the title of the chart to “Average engine oil, tire pressure, speed and maintenance probability by model”.

Click the blank area to add new visualization:

1. Select Scatter Chart visualization from visualizations.
2. Drag the Model field into the Details and Legend area.
3. Drag the fuel field into the X-Axis area, change the aggregation to Average.
4. Drag engineTemparature into Y-Axis area, change the aggregation to Average
5. Drag the vin field into the Size area.

Change the chart Title to “Averages of Fuel, Engine Temperature by Model”.

The final report will look like as shown below.

Pin visualizations from the reports to the real-time dashboard

Create a blank dashboard by clicking on the plus icon next to Dashboards. You can name it “Vehicle Telemetry Analytics Dashboard”

Pin the visualization from the above reports to the dashboard.

The dashboard should look as follows when all the three reports are created and the corresponding visualizations are pinned to the dashboard. If you have not created all the reports, your dashboard could look different.

Congratulations! You have successfully created the real-time dashboard. As you continue to execute CarEventGenerator.exe and RealtimeDashboardApp.exe, you should see live updates on the dashboard. It should take about 10 to 15 minutes to complete the following steps.

Setup Power BI batch processing dashboard

Note

It takes about two hours (from the successful completion of the deployment) for the end to end batch processing pipeline to finish execution and process a year worth of generated data. So wait for the processing to finish before proceeding with the next steps.

Download the Power BI designer file

• A pre-configured Power BI designer file is included as part of the deployment
• Click the Power BI node on the diagram view and click Download the Power BI designer file link on the properties pane
• Save locally

Configure Power BI reports

• Open the designer file ‘VehicleTelemetryAnalytics - Desktop Report.pbix’ using Power BI Desktop. If you do not already have, install the Power BI Desktop from Power BI Desktop install.
• Click the Edit Queries.

• Double-click the Source.

• Update Server connection string with the Azure SQL server that got provisioned as part of the deployment. Click the Azure SQL node on the diagram and view the server name of the properties pane.

• Leave Database as connectedcar.

• Click OK.
• You will see Windows credential tab selected by default, change it to Database credentials by clicking on Database tab at right.
• Provide the Username and Password of your Azure SQL Database that was specified during its deployment setup.

• Click Connect
• Repeat the above steps for each of the three remaining queries present at right pane, and then update the data source connection details.
• Click Close and Load. Power BI Desktop file datasets are connected to SQL Azure Database tables.
• Close Power BI Desktop file.

• Click Save button to save the changes.

You have now configured all the reports corresponding to the batch processing path in the solution.

Upload to powerbi.com

1. Navigate to the Power BI web portal at http://powerbi.com and login.
2. Click Get Data
3. Upload the Power BI Desktop File.
4. To upload, click Get Data -> Files Get -> Local file
5. Navigate to the “VehicleTelemetryAnalytics – Desktop Report.pbix”
6. Once the file is uploaded, you will be navigated back to your Power BI work space.

A dataset, report and a blank dashboard will be created for you.

Pin charts to the existing dashboard Vehicle Telemetry Analytics Dashboard in Power BI. Click the blank dashboard created above and then navigate to the Reports section click the newly uploaded report.

Note the report has six pages:
Page 1: Vehicle density
Page 2: Real-time vehicle health
Page 3: Aggressively Driven Vehicles
Page 4: Recalled vehicles
Page 5: Fuel Efficiently Driven Vehicles
Page 6: Contoso Logo

From Page 3, pin the following:

1. Count of VIN
   
2. Aggressively driven vehicles by model – Waterfall chart
   

From Page 5, pin the following:

1. Count of vin
   
2. Fuel efficient vehicles by model: Clustered column chart
   

From Page 4, pin the following:

1. Count of vin
   
2. Recalled vehicles by city, model: Treemap
   

From Page 6, pin the following:

1. Contoso Motors logo
   

Organize the dashboard

1. Navigate to the dashboard
2. Hover over each chart and rename it based on the naming provided in the complete dashboard image below. Also move the charts around to look like the dashboard below.
   
   
3. If you have created all the reports as mentioned in this document, the final completed dashboard should look like the following figure.

Congratulations! You have successfully created the reports and the dashboard to gain real-time, predictive and batch insights on vehicle health and driving habits.