Innovation Lab Partners

Over the last five years, QMSS has been incredibly fortunate to do research alongside an amazing set of research partners (see partial list below).  

In each case, QMSS student teams have brought the breadth and depth of their analytical expertise to practicum projects in the Innovation Lab.  We are proud that many research partner organizations find so much value in these engagements, that they sign up to do another research project again in the future.

Across a wide range of projects, QMSS teams have designed predictive models, built interactive dashboards, and created data-driven tools to inform policy and guide organizational decision-making.  They have analyzed behavioral and survey data, explored applications of generative AI, evaluated algorithmic fairness, optimized customer experiences, and refined models for medical and environmental forecasting.  

Whether developing prototypes, uncovering insights from complex datasets, or translating research into actionable strategies, QMSS student teams have continued to show how data science and social inquiry together can drive smarter and more effective outcomes.  

Partial List of Innovation Lab Research Partners (2020-2025)

Consulting & Analytics Firms:
Blue Labs, Boston Consulting Group, Ekimetrics, IBM Consulting, KPMG, Lovelytics, Quantilope

Academic, Research & Think Tanks:
Columbia Climate School, Joseph C. Cornwall Center for Metropolitan Studies at Rutgers University–Newark

Government & International Organizations:
Department of Licensing and Consumer Protection, The Census Bureau-The Opportunity Project, UN Development Programme, UN Refugee Agency

Nonprofits & Advocacy Organizations:
Just Capital, Natural Resources Defense Council, New York Common Pantry, Memorial Sloan Kettering Cancer Center

Startups & Private Sector Companies:
.406 Ventures, AltSurya, Lyssle, Priceline, The Black List

 

Representative Practicum Project Details:

UN Joint SDG Fund

Problem: The UN needed to understand how the 17 Sustainable Development Goals (SDGs) interact, in order to better track progress and allocate resources.
Solution: We created two models: a text-based network analysis using Voluntary National Review documents, and a numerical linkage model using the SDG Indicator Database (231 indicators, 169 targets).
Result: The models uncovered specific groups of goals—like health (SDG 3), education (SDG 4), and inequality (SDG 10)—that are tightly linked in practice. This insight helps policymakers target actions that drive progress across multiple goals at once, making development efforts more efficient and impactful.

Lovelytics (Customer Segmentation)

Problem: The client needed to identify customer segments and tailor marketing strategies using credit card and household data.
Solution: We applied K-means and bisecting K-means clustering on transaction records (2019–2021) and mapped them to demographic/psychographic variables from Epsilon.
Result: We found two distinct customer types with clear behavioral and demographic differences. One group, while smaller, consistently spent more—especially in categories like supermarkets and home improvement. These insights gave the company a data-backed roadmap for prioritizing high-value customer outreach and refining their marketing spend.

The Black List (Script Insights)

Problem: With thousands of scripts and reviews, The Black List needed tools to extract trends and improve discoverability.
Solution: We performed tag co-occurrence analysis, built a script similarity classifier using content metadata (genre, roles, tone), and visualized diversity trends in gender and race across years.
Result: Our analysis revealed meaningful patterns in script content and reader feedback, including which tags tend to co-occur and where diversity gaps appear. We also developed a tool to group similar scripts by genre, roles, and tone, giving producers and writers a faster, smarter way to discover compelling content.

The Opportunity Project (US Census)

Problem: Rural towns lacked accessible tools to use federal datasets for economic and environmental planning.
Solution: We co-designed a prototype (“R Story”) with local leaders in Manistee, MI using publicly available EPA, Census, and USDA data. The tool includes custom dashboards tailored to local entrepreneurs, residents, and developers.
Result: By creating a one-stop data tool for rural leaders, we made it dramatically easier to find and use critical planning data. The tool is already supporting leaders in Manistee, Michigan as they make smarter redevelopment decisions and apply for funding—proving that small communities can fully leverage federal data when it’s accessible and actionable.

KPMG (COVID Forecasting)

Problem: KPMG needed dynamic forecasting to prepare for region-specific COVID disruptions.
Solution: We built a set of SEIRD-based predictive models using data from Johns Hopkins, Google Mobility, Oxford policy trackers, and the US Census, with region-level customizations for all 50 states and largest cities.
Result: Our models generated accurate, region-specific forecasts that helped KPMG simulate COVID scenarios for cities and states. These insights gave organizations a clearer picture of potential disruptions—allowing them to adapt business and financial plans more confidently in a fast-moving environment.

If you are interested in submitting an initial proposal, please use this form to start a conversation with us. Potential Partnership Proposal