Matching Algorithms

1.0 Project Requirements

The client, Erika Bailey, is an International Transition Advisor at the Center for International Experience (CIE) at the University of Toronto (UofT) [1] and the coordinator of the UofT mentoring program, iConnect [2]. The iConnect International Mentorship Program is funded by the CIE and is responsible for facilitating the cultural transition of over 150 mentees by matching them with appropriate mentors and promoting intercultural dialogue [2]. The client finds the current matching system inefficient and seeks a solution.

1.1. Problem Statement

The client’s problem is rooted in the acquisition of mentor/mentee information and the processing time for creating a match[3]. Matching mentors and mentees is a challenging and crucial step of a mentorship program; a productive mentoring relationship relies on an appropriate match[4]. The iConnect system comprises of three levels; the mentor and mentee relationship, the leader who oversee these relationships [5] and the client, Erika Bailey[6]. Mentees can apply from May to September; ideally the matches are made before their arrival in Canada (ie. during the summer) [2]. Two people are responsible for the match-making process; a work-study student and the client [3]. There are 30 mentors and 150 mentees.

Communication with prospective mentees is done via email, with at least one email for every student involved. Matches are done manually after analyzing data stored in a spreadsheet [3]. Up to 2 hours of work per application is required[3]. The client has only 8 hours a week to dedicate to this program, which makes the current system inefficient[3]. Moreover, manual matching is not recommended for mentorship programs consisting of over 100 people, since the margin of human error is likely to increase[7]. The high amount of information and emails the client manually processes make the system labour-intensive and tiring.

The client requires a more efficient method of acquiring and...