Promising Safer Roads with Business Intelligence
How does one ensure that an organization’s fleet of more than 5,000 vehicles continue their operations safely and incident-free? For one of our clients, we achieved this by developing a two-tier Business Intelligence solution. The frontend tier leveraged MicroStrategy’s direct database access and in-memory processing capabilities with a real-time Azure Data Lake at the backend that incremented up to 5 million records a day.
Our consultants built the solution keeping the following objectives in mind:
- Ensure vehicle safety
- Reduce road-related incidents
- Protecting the driver and encouraging responsible behavior
- Assist risk reduction
- Improve operational efficiency
The MicroStrategy dashboards allowed the fleet management team to remotely monitor shifts in the metrics, and identify trends in driving behavior. Aided by these BI dashboards, users were able to access broader views of the fleet across numerous business partners and give external management access to drivers’ performance. Fleet management has been an ongoing activity for this qordata client. Thanks to our solution, their data analysis activity has been reduced from weeks down to hours.
This project adds value with exciting, and previously unavailable features in the vehicle safety area. BI reporting now includes spatial event visualizations that allow business users to analyze vehicle movement, events and violations that earlier solutions did not cover. The option of ad hoc analysis per location also exists now, as well as early detection and awareness of potential hot zones.
Solution Benefits
- > 7x reduction in fleet analysis time
- Preemptive incident flagging
New Features
- Spatial event visualizations
- New vehicle movement, event and potential violation view capture
- Ad hoc analysis
- Early detection
- Flagging potential hot zones
Road Safety
For Veraqor’s BI consultants, assimilating publicly available datasets on traffic statistics and GPS was easy. For the most part, the data was structured, authenticated and periodically published. Together with real-time changes in longitude and latitude, the challenge was to pair this with the unstructured data collected from sensors (such as dashboard cameras) within the client’s fleet.The fleet’s sensors captured everything from images of the driver’s seatbelt, dashboard data to ascertain speed limits, and even images of the driver to make sure that he/she was physically fit to drive.
Capturing data on to the client’s cloud server was not difficult. Although data from each source (the vehicle tracker, the cameras) belonged to different vendors, and was being collected in separate repositories.
The resulting deluge of high data volumes—up to 5 million transactions per hour, increased processing time significantly. It also led to a higher incidence of anomalies, or dirty data. Because these factors prevented data visualization from full utilization, HSE managers and other decision-makers within the client’s organization struggled to get a completely accurate picture of their fleet’s road performance.
Engaging Veraqor’s Business Intelligence Consultants
The client had already engaged Veraqor’s consultants, and was familiar with their expertise in MicroStrategy and other BI tools, especially the Microsoft BI stack. Based on experiences in earlier projects, the client expected high service levels, quality consistency and timely delivery of project milestones from the Veraqor team.
Multi-Component BI Infrastructure
So what could offshore BI experts do to prevent road accidents in the client’s home country?
With the right data skills, almost everything. The Veraqor team proposed a multi-component solution, starting with the portal and leading to five interface(s).
Each interface (screen), captured data (historical or real-time), from the client’s own repositories or publicly available databases, and used these to generate insights to help either prevent or diagnose a traffic incident.
Geo-Spatial Visualization
This screen visually represented both historical and real-time data related to heat maps and road utilization. Using this, users could estimate ETAs and riskiness per route, propose alternative itineraries and more. Users also benefited from an interactive filter option, whereby they could pinpoint incidents by time, location, contractor and even vehicle type.
Event Detection
This screen combined data from recorded and detected events to help release warnings about concerns that could escalate into incidents. For instance, it represented the occurrence of speeding, overtaking, violating road signs and unskilled driving.
Supporting Applications
To enrich output, and to provide as comprehensive a view as possible, the Veraqor solution captured data from internal repositories as well. These databases provided input on road conditions, speed limits, overtaking history and even vehicle records. More detailed configurations became possible after the solution took in data about road signage, training records, the client’s database of contractors and even cars in proximity.
Business Intelligence
During needs assessment, the Veraqor team identified issues in the existing model. In the ETL Stage, the source system was producing numerous duplicate records, which was fixed by Veraqor data experts. Volume management was also becoming problematic because there were two separate fields for the same identifier. In the Veraqor solution, these identifier keys were combined into a single unique primary key column with the help of a mathematical formula.
The solution also saved on an additional data transformation cycle by applying transformations in the new data lake. Further changes were made during Data Structuring, Profiling and Dimension Modelling.
Notifications
From a business-user perspective, solution notifications provided the handiest and most user-friendly access to insights. The subscription feature allowed users to customize reports by events, vehicles, locations and even time of incident. This made data management at the user’s end both practical and efficient, as it limited view to need, and provided accessibility by email and text (SMS) message.
Using Analytics to Complete the BI Puzzle
As mentioned earlier, users in the client’s organization already had data on hand about road safety. New insights emerged by adding the number of traffic violations that would increase the possibility of accidents. This helped in two specific ways: Closing the loop meant that potential issues could be identified proactively, therefore forewarning drivers and/or providing them with feedback, real-time alerts and notifications.
The interactive BI component equipped users with new insight tools, and automated the reporting process significantly.
Finally, it offered predictive analytics on driver fatigue and traffic congestion, preemptively addressing issues that could otherwise be fatal.
Intelligent Insights, Safer Roads
The objective of the this project had been clear from the start, and was far from unique. Public and private sector organizations across the world had been undertaking initiatives to prevent road-related casualties.
What was new about Veraqor’s solution is that it settled the debate on how road-related incidents should be approached. Instead of defining incidents by static factors like geographic proximity, vehicle type et al, it made use of real-time data to provide a 360o-view to ensure that no possibility of a traffic accident was left unmitigated or uncontrolled.