Main Content

TD MDAL Research Grant

Topics using frontier methods in data analytics, machine learning, or text mining as applied to a variety of business problems were welcome.

This year marked the inaugural call for proposals for the TD MDAL Research Grant. Data analytics is relevant across a wide range of areas, applications from faculty and PhD students from all areas of the Graduate Department of Management were considered. Grants of up to $10,000 were awarded for research costs related to data analytics projects. 

The lab is proud to award funds to the successful projects listed below. 

Faculty and PhD Student recipients include:

Daniel GoetzAssistant Professor of Marketing“Learning from Even Weaker Ties: Peer Effects from Strangers in Online Purchases” (with Wei Lu)

Project description:

This project aims to measure the relative strength of peer effects from friends versus peer effects from randomly encountered strangers in the context of consumers’ product adoption decisions. Our goal is to answer whether peer effects in product adoption primarily capture the effect of mere additional exposure to the product, or whether other information about the product is conveyed that depends on the nature of the peers' relationship.

Andrey GolubovAssistant Professor of Finance,“Are takeovers costly for workers?”

Project description: 

Using newly available administrative data on the universe of Canadian firms and individuals from Statistics Canada, this project will investigate whether takeovers are costly to target firm workers in terms of jobs, earnings from employment, and overall incomes after taking into account any substitute sources of pay.

Sheng LiuAssistant Professor of Operations Management and Statistics, “Smart Urban Transport and Logistics Empowered by Data Analytics

Project description: 

New technologies and innovative business models are leading to connected, shared, autonomous, and electric solutions for the tomorrow of urban transport and logistics. A tremendous amount of data is generated every day from public transit operators, mobility service platforms, and logistics service providers. This research program develops data-driven solutions to emerging urban transport and logistics challenges, including multi-modal transit planning, last-mile delivery, and urban warehouse fulfillment.

Matthew OsborneAssociate Professor of Marketing“Interactive Customer Feedback in the Digital Economy: When and How Creators Should Respond to Customers?” (with Minjee Sun)

Project description:

Technological developments have enabled content producers and their customers to interact with each other in real time. Moreover, content producers may modify their offerings in response to customer feedback. Using data from an online book platform, we explore how novel writers modify their novels as a result of customer feedback and what kind of modifications enhance the novels’ performances. In this research, we apply natural language processing and text mining to characterize readers’ comments and quantify the changes in writers’ novel contents. Our preliminary analysis suggests that the sentiment of readers’ comments matters, as well as writers’ past publishing experience.

Shreyas SekarAssistant Professor of Operations Management“Learning to Combat Fraudulent Behaviour in E-Commerce"

Project description: 

Digital marketplaces such as Amazon, AirBnB, and Upwork play a significant role in influencing our choices by means of recommendation and ranking algorithms. As a result, sellers on these platforms are incentivized to resort to fraudulent behaviour---e.g., fake reviews, click farms, duplicitous listings, bots, etc----to game the algorithm. Existing methods to combat fraud tend to be reactive rather than proactive. This project will centre around developing prescriptive analytics and proactive learning algorithms that converge to socially desirable outcomes despite the presence of manipulative behaviour.

Avni ShahAssistant Professor of Marketing“Payday Lending and Its Implications: Designing More Optimal Lending and Repayment Strategies (with Dinara Akchurina and Andre Cire)

Project description:

Payday loans are considered an expensive way for consumer to borrow money, and yet the number of Canadians using payday loans has doubled over the last decade. For many borrowers, failing to repay the amount borrowed can result in far more costly expenses as the interest payments can result in repayment amounts that are nearly six times the original borrowed amount. In this research project, we use a rich dataset combining precise borrower-specific characteristics, outstanding loan-specific characteristics (loan amount, loan purpose, interest rate, loan duration, loan type—i.e., cash advance or title loan, repayment schedule) coupled with a detailed list of banking transactions to investigate a number of important research questions: What drives consumers to take out a payday loan in the first place? What trade-offs are made within the pay cycle or what external forces drive this need? When (and for whom) is it financially prudent to use payday loans rather than other sources of short-term credit? What factors increase the likelihood of timely payday repayments? Do reminders via text or via direct phone contact increase the likelihood of paying a loan on time? Using insights from our data and survey responses, we will develop novel predictive, prescriptive, and operational models that can improve consumer decision-making and product design.

Mengze ShiEllison Professor of Marketing, “How to Engage the Online Community?A Textual Analysis of Comments” (with Clarice Yulai Zhao)

Project description:

Digital platforms enable a massive number of independent creators to engage their online communities. The creators facilitate and participate in the discussions about their contents. The proposed project intends to study the nature of engagement activities and the impact on the strength of customer relation, using purchase and comment data from a major online publishing platform in Asia. We plan to use natural language processing methods to analyze the unstructured data, convert them into quantifiable measures, and relate them to behavioral outcomes. We expect the research outcomes to inform the effective engagement strategies in the online communities.

Claire TsaiAssociate Professor of Marketing, “Paying to Write Good Reviews? Positive Skewness of Distribution of Consumer Reviews on E-Commerce Platforms vs. Review Sites” (with Ying Zeng, and Wei Lu)

Project description:

This research applies econometrics, machine learning, and text analysis approaches to investigate whether and how consumer review distributions differ across different types of web platforms, and the mechanisms driving these differences. Are consumers more critical on review websites than on e-commerce platforms? If this is the case, what might be driving this difference? While existing research has shown that the distribution of consumer reviews is skewed, understanding why the degree of skewness varies across websites remains largely unknown. Our research fills this gap by using data analytics and big data to study this question.

Claire TsaiAssociate Professor of Marketing, Good Things Satiate and Bad Things Escalate?’ Consumers Adapt to Positive Experiences and Sensitize to Negative Experiences” (with Kailuo Liu)

Project description: Forthcoming.


Irene YiAssistant Professor of Finance“Which Firms Require More Governance? Evidence from Mutual Funds' Revealed Preferences”

Project description:

This project estimates mutual funds’ preferences for corporate governance structures, by examining how mutual funds voted on companies’ shareholder proposals. The project develops rankings that measure which firms benefit more from adopting governance provisions that increase shareholder rights, from the perspective of mutual funds. In doing so, the project implements a novel machine-learning algorithm: the Metropolis-Hastings Markov chain Monte Carlo algorithm of Vitelli et al. (2018). The project complements the literature that questions the “one-size-fits-all” approach toward governance.

Rolando Campusano, PhD Student, Economic Analysis and Policy"Delineating Neighborhoods using Location Choices

Project description:

Research using neighbourhoods as the unit of analysis has relied on administrative definitions that have been delineated from a process that does not coincide with agents’ decision problems. This produces a spatial misalignment between administrative and "economic" boundaries that bias research findings and the policies designed around them. I propose a novel methodology to delineating neighbourhoods based on a machine learning algorithm that groups locations based on revealed preferences. I apply it to delineate Toronto's industrial and residential neighbourhoods and show that they are not like their administrative counterparts. Neighbourhoods are different across industries or property types, have an elliptical shape and tend to locate around major streets.

Stacey Choy, PhD Student, Accounting, Inside the Black Box of Private Communications: Evidence from Taxi Ride Patterns between Managers and Analysts in New York City” (with Ole-Kristian Hope)

Project description:

This study constructs a novel measure that aims to capture face-to-face private communications between firm managers and sell-side analysts by mapping detailed, large-volume taxi trip records from New York City to the GPS coordinates of companies and brokerages. Consistent with earnings releases prompting needs for private communications, we observe that daily taxi ride volumes between companies and brokerages increase significantly around earnings announcement dates (EAD) and reach their peak on EAD. After controlling for an extensive set of fixed effects (firm-quarter, analyst, and year) and other potential confounding factors, we find that taxi rides undertaken around EAD are negatively associated with analysts’ earnings forecast errors in periods after EAD. Analysts having more taxi trips around EAD also issue more profitable recommendations after EAD. Our results suggest that analysts may obtain a private source of information orthogonal to their pre-existing information from these in-person meetings, which may help them better understand the implications of current earnings signals for future earnings.

Mohsen Foroughifar, PhD Student, Marketing,How Do Algorithmic Price Suggestions Impact Home-Sharing Markets? Evidence from Airbnb"

Project description: 

Starting from December 2015, Airbnb used machine learning to recommend "smart prices" to its hosts with the hope that they use these recommendations in their pricing decisions. Smart Pricing algorithm uses past data to recommend product-specific prices to Airbnb hosts. Although there might be some benefits with using smart pricing algorithm - being free, easy to turn on and off, and allowing hosts to set price boundaries - many hosts do not use it.  In this work, we study the impact of these algorithmic price suggestions on home-sharing markets. We examine how the introduction of Smart Pricing has impacted Airbnb's hosts and what the subsequent welfare implications of this technology are on the home-sharing markets.

Zheng Gong, PhD Student, Economic Analysis and Policy, “How does popularity information affect product design?” (with Guangrui Li, York University)

Project description: 

This project aims to examine the impact of popularity information revealing on firms’ product design strategy. We make use of a policy change – that articles’ popularity is revealed to subscribers – on Wechat Official Account platform, which is the biggest blog platform in China. We are interested in applying natural language processing technics to understand the following aspects of product design: first, the quality and amount of advertisement in the contents; second, the topic choice of the contents; and the tendency to use click-baits in the titles.

Saman Lagzi, PhD Student, Operations Management“Model-Free Assortment and Bundle Pricing with Transaction Data

Project description: Forthcoming.


Marco Salerno, PhD Student, Finance“Who Should Buy Stable Firms?”

Project description:

The current low interest rate environment poses challenges for liability-driven investors (i.e., pension funds): while being a good hedge towards liabilities, long-term bonds do not provide enough income to repay investors’ liabilities. This project shows both theoretically and empirically that liability-driven investors should tilt their portfolios towards “stable equities”, which are defined as firms that operate in stable sectors of the economy. A new methodology has been developed to classify stable equities using textual analysis and detailed data on consumption by type of product.

Mingyue Zhang, PhD Student, Accounting“Determinants and consequences of human capital management disclosure”

Project description: Forthcoming.