Solving real-world problems through online discourse

In April of 2021, an Edmonton and Vancouver-based company called Stantec received a contract to begin testing for a project in collaboration with Transport Canada. The project focuses on using one of Stantec’s online services, Customer Sentiment Analytics, to help the Government of Canada address the impacts of disruptions in the transportation system.

Stantec’s Customer Sentiment Analytics gathered data from engagement tools like surveys, social media, and online news and looked for trends and patterns on transportation issues. These issues can range from traffic on a busy highway, to a just-burst pipe, to faulty traffic lights.

By using artificial intelligence, the service can also detect the public’s views on certain issues, and how fast those views change over time. With this, the government can uncover which operations and services are causing shifts in public sentiment and the reasons behind those shifts.

Stantec’s online service provides the Government of Canada with a wide overview of community needs on transportation, allowing us to address these needs in the most efficient and cost-effective ways.

The technology is installed on a secured cloud network with 4 main parts:

  • data fetcher
  • data processing
  • data storage, and
  • a data visualization dashboard

The data fetcher connects to publicly available data sources, like social media, online surveys, and the news. It looks at information on public engagement, mentions, keywords or phrases, and geographic locations. This data then moves to the data processing area.

In data processing, the system uses artificial intelligence (AI)-based natural language processing, text analytics, and other forms of data analysis to process the information. This is where online public sentiments on government operations are labelled “positive”, “negative”, or “neutral”.

The AI also groups issues that are brought up online, like lack of water, road closures, or power outages. It can also detect the location of these issues, as well as key online advocates for addressing these issues. Additionally, the algorithm can predict any service disruptions that will occur in the future.

Once processed, this data will be moved to the data storage area, which keeps the data in a secured Microsoft Azure Environment hosted within Canada. The data can be viewed via computer on the data visualization dashboard. Critical alarms for urgent issues and major events are highlighted on the dashboard, and can be sent via email to federal departments. The data can also be exported as an Excel download, and to client systems through a secure Application Programming Interface (API) connection.

The system’s goal is to help the Government of Canada understand why Canadians feel certain ways about issues and address those feelings. In testing the abilities of Customer Sentiment Analytics, the team hoped to find an accurate and efficient way to address transportation-related issues before they got out of control.

Testing the effectiveness of Stantec’s Customer Sentiment Analytics ended in November of 2021 with positive results. The technology was great for data visualization, and all key performance indicators (KPI’s) were either met or exceeded. Some examples of transportation-related issues that were discovered and addressed using Stantec’s technology were:

  • congestion at the Port of Vancouver, and
  • wildfire hazards near Vancouver rail lines

As work continues, the team at Stantec and Transport Canada hope to give the software the ability to filter out data from government social media accounts. This because the most active transportation-related accounts are government accounts. Including this data could distort the results of an analysis.