About Scius

Scius is a British Columbia based research and consulting firm focused on the emerging construction technologies and materials industry. Their services include market insights, policy and program consulting as well as building project and technologies research.

The Problem

In order to support their research and consulting efforts, Scius regularly engages in significant data gathering efforts. As Scius worked primarily on a project by project basis they were encountering challenges in leveraging the results of past data gathering to support new projects. Knowledge of what existed was spread amongst the company’s researchers, meaning that efforts were at risk of being duplicated due to low company wide data visibility. Scius additionally wanted to expand their business model into the data product space. To support the efficiency of their current business and expansion goals, the company needed a low barrier to entry system allowing researchers to access, analyze and add to the company’s knowledge base.

The Solution

Tensory has a strong background in data integration and knowledge management, leveraging years of experience in the government data sector. To support Scius’ goals, Tensory delivered the following:

• Built and promoted a culture of data sharing and standardization across the organization. Prior to Tensory’s involvement, each researcher would create their own data schema on a project by project basis. A data collection guide providing standards into column header formats, data types and best practices ensured that researchers maintained flexibility during their work while ensuring the easy integration of new data down the road;
• Set up a cloud hosted web platform that allowed researchers to add new data sources, search through previously collected data and export relevant datasets. Data was stored using a graph database, providing a flexible and extensible schema while maintaining an easy to understand structure;
• Developed a machine learning driven data deduplication layer used to link instances found in multiple data sources. For example, if the province of BC is mentioned in one data source while another refers to it as British Columbia, the system was capable of identifying the two as the same instance and then linking them. This increased the reliability of the knowledge base as a source for insights and analysis.
• Developed a natural language query interface leveraging Large Language Models to allow users to analyze data without having to learn technologies such as the Cypher Graph Query Language. Users could prompt the system with a question and receive an answer based on the knowledge base. The results could then be downloaded for delivery to clients and additional internal analysis;

Results & Benefits

Since the deployment of the system, Scius has been able to systematically add and monetize their data at scale. New researchers need little training to be able to leverage the company’s past projects and knowledge. New data sources can be added without the need for manual reconciliation with current data, increasing the breadth and depth of data that Scius can use.

‍

"Choosing to parter with Tensory was a great decision for our business. The bespoke application developed by Julien has helped transform the knowledge and data capture process within our organization. The solution has been extremely effective, improving our operational efficiency by making data more accessible. We were impressed by Tensory's innovative approach and commitment to understanding our needs."
_{― Mark Devereux, Principal, Managing Director at Scius}

‍