Montreal, January 14, 2021
“The McGill MMA Analytics Consulting initiative is an 8-10 month Experiential Learning engagement where we aim to build stronger ties between Industry and Academia by working side-by-side in trying to solve today’s most pressing business problems using the latest Data Science & Technology solutions. Teams of 4-6 MMA students with complementary skills are matched, then paired with industry partner-clients to try to find unique solutions for critical issues within their organizations and Industries.” – McGill University, September 2020
Like most recruiting and selection processes, the partnership between Euler Hermes Canada (EHCA) and McGill University was established through a contact. In late 2018, an associate at McGill’s student placing area connected with Euler Hermes Canada’s CEO, David Dienesch, about the MMA consulting initiative. She knew the company managed great amount of data and believed it would be a good fit for the program.
Project Description (from McGill's website)
In this project, students collaborated with the Credit Insurance team on developing a solution to extract data from financial reports in a more automated fashion and investigate risk assessment factors. The provided recommendation was created for standard report readers who utilize text analytics to leverage new sources of risk factors to increase coverages and decrease Net Loss Ratios.
- Fairly manual process on the data extraction phase
- Utilizing traditional risk factors based on past structured data
- Financial Statement reports
- Dun & Bradstreet reports, history, comments
- Historical Risk Gradings
- Leveraging sentiment analysis on public data (i.e. analyst reports) to look for risk factor applicability and correlations
- Use of Gunning Fog Scores for Sector and industry grading
- Testing and Model validation using Logistic Regression, SVM and Random Forest algorithms
Technical Solutions Developed
- Data sources included 10K filings and Twitter analysis
- Text extraction involved identifying scanned PDFs, correcting for skewness, extracting text from images, and placement of correct text using ML algorithms on the extracted words
A Phenomenal Experience
EHCA submitted four issues internal departments faced on a daily basis. One of the use cases submitted, and ultimately the one that was chosen, was the pressing need to automate the spread of financials – a tedious and manual process not just for Euler Hermes, but also for many financial organizations.
The McGill team was composed of four students, Ruben Alba, Business Strategist; Younes Amor, Solution Architect; Cindy Qu, UX/UI; and Tamar Bulka, Modeler.
During the 8-month project, both parties met regularly and developed a pleasant and fruitful working relationship. It was truly a phenomenal experience. Euler Hermes’ employees involved in the project developed a different set of skills and curiously enough, got to see their company through a different lens. They had the opportunity to put themselves in the students’ shoes and learn how people outside view and understand the company with little to no experience in the trade credit insurance industry.
The students delivered on the use case by thinking outside the box and designing the best solution for the problem. They successfully used optical character recognition (OCR) and sentiment analysis to automate the task of grading financial statements. Impressed by the outcome, the Canadian team sent details of the project to Head Office in Paris, where data scientists will integrate the algorithms with the company’s own OCR projects!
In addition, the team at Euler Hermes is delighted to continue to benefit from the students’ talent. Following the project completion, Human Resources extended an employment offer to Tamar Bulka, who is now Regional RICC Controller, Information & Grading.
This journey required a lot of group effort and the teams at EHCA were bold at pushing boundaries. As McGill University mentioned in their Partner-Client of the year letter, Euler Hermes Canada stood out because they successfully created a comfortable learning environment for the students, and determined highly valuable and thought-provoking use cases, while consistently challenging student consultants to push the envelope to strive for more in their pursuit of deeper learning and skills development.
It is also worth mentioning that the MMA student consultants achieved one of the top marks amongst other projects in the course, and as the letter states: This could only be achieved through strong commitment and dedication to excellence by all people involved.
More to Come
The success of the first project has given EHCA a second opportunity at working with McGill University MMA students. This year’s project consists of evaluating Canadian SMEs’ propensity to buy trade credit insurance.