If you are a researcher, academic, or data-driven thinker, the LinkedIn Economic Graph Research Program (EGRP) is your opportunity to team with LinkedIn to solve some of the most challenging economic problems of our time.
Today, I am happy to provide an update on the new research teams that have joined the EGRP on an ongoing basis, and to announce a new call for proposals for new teams who are looking to pursue challenging questions in the fields of economics, analytics, or artificial intelligence, in partnership with LinkedIn. Teams who are selected to participate in the EGRP will have the opportunity to pursue their research alongside our experts in these fields and access de-identified or aggregate data from LinkedIn in a way that respects the privacy of LinkedIn members.
Asking important questions
LinkedIn is entrusted with a unique dataset and believes that this data should be used in service of creating economic opportunity for every member of the global workforce. Our plan is to achieve this through the creation of the Economic Graph, a digital mapping of the global economy, which allows us to connect talent with opportunity at massive scale. This highly-structured graph provides rare insight into the functioning of the global economy and has been leveraged at various times by organizations such as the World Economic Forum, the World Bank, state/local governments, and nonprofits to help identify macroeconomic trends and opportunities. Following last year’s call for proposals, several new research teams joined the EGRP as partners, exploring new and compelling research projects in their fields using de-identified or aggregate data from LinkedIn, such as:
Researchers Rossano Schifanella from University of Turin, Nicola Perra from Greenwich University, Indiana University professor Alessandro Flammini, and collaborators are looking at micro and macro teams composition and the performance of the businesses the teams work for. For example, what are the impacts of having diverse skills and roles on businesses, from startups to larger and more well-established firms? The team plans to adopt a blending of network science and machine learning to model team intelligence at a global scale.
Researchers from Harvard Business School and the USC Marshall School of Business will be exploring the correlation between firm success and the connections and backgrounds of team members. This research may help reduce organizational conflicts and help entrepreneurs and established firms to improve team formation, based on LinkedIn data.
Researchers from University of Illinois Urbana–Champaign are exploring a variety of topics related to better understanding the semantics of links in networks and the relationships embodied within them. For example, the team looks to analyze how manager-employee relationships are different from peer relationships within the same organization and the implications of these different relationships for the relevance of content on LinkedIn. The team notes that, despite the rich research on the semantics of nodes, there has been less study of the semantics of links in networks.
These teams join existing Economic Graph Research teams from schools such as Indiana University (where the team is algorithmically identifying “micro-industries” based on workers’ firm-to-firm transitions) and MIT (which is using LinkedIn Economic Graph data as well as financial data to test theories about investment in human capital). Using our Economic Graph data (including de-identified data), these teams have already captured real-world economic phenomena at a level of detail and scale that rivals best-of-breed methodologies from agencies like the U.S. Bureau of Labor Statistics.
Unique ongoing research into analytics, economics, and AI
While our own researchers have made great strides in the application of this data to addressing both real-world problems and some interesting theoretical questions, we believe that partnering with academic researchers has tangible benefits, not only for LinkedIn members and customers, but for the global research community as well. Consider the following:
With many academics experiencing difficulty getting access to novel, real-world datasets, our EGR program provides researchers with an opportunity to test or validate their theories on one of the largest and most robust datasets of professional and economic networks.
Outside researchers often ask questions and bring perspectives that may not be already present within our own research team. By having more insights from different perspectives, we can address questions that would be overlooked if we were to rely solely on the perspectives from within our own team.
Everyone, from the engineering teams building LinkedIn to the members and customers using our platform, has unconscious biases that shape their view of the world. Uncovering and addressing these biases is hard to do, but every insight can lead to important, positive changes in the way that we make product and design decisions at LinkedIn. For example, an earlier Economic Graph research project provided useful insight into self-representation differences in member profiles, helping us build better recommendation systems. Recently-published research from Tufts University used A/B test data from LinkedIn to show how information cues can be used to avoid unconscious bias in job application behavior. These research insights both answer questions about human behavior that can be applied in other industries and provide those of us at LinkedIn with information that can help us develop better products for our members.
“Members First” data access
We have a policy of “members first,” which means that we must respect the privacy and ensure the security of our members’ data by honoring their settings and privacy choices. In any research we conduct, we seek to ensure that the purpose of the research does no harm to our members, is done in accordance with ethical and legal standards, and minimizes the amount of data that is accessed and used. For example, in EGRP we only permit access to aggregated or de-identified data and only within a secure “sandbox” environment. We do not permit access to data regarding the private actions of our members. Accordingly, researchers’ proposals and requests to access data will additionally be vetted by LinkedIn’s legal and security teams for privacy and security concerns.
Members first also means that the research project must relate to economic opportunity with an eye toward enabling a level playing field for economic outcomes. Research proposals and publications are peer-reviewed by experts in fields including, but not limited to, analytics, AI, and economics. In selecting proposals, we look at a few key factors, such as the novelty of the research idea, the data availability and accessibility, the team’s commitment, and the duration of the research.
Call for proposals
Today, I would like to announce that interested researchers from universities, think tanks, and other non-profit entities can download the proposal template and submit their proposals for the Economic Graph Research Program until Dec. 1.
Proposals for the EGRP are being solicited in three broad topic areas:
Analytics: At LinkedIn, our mission is to drive understanding and impactful decision-making through rigorous use of data. Our analytics are deeply tied to core modules of our ecosystem, including product, marketing, and sales, to name a few. We are looking for research proposals that leverage big data analytics and data science to understand relationships in the Economic Graph, such as: skill set trends and predictions, talent supply and demand gaps (at micro and macro levels), and the economics of professional mobility and talent migrations.
Economics: We aim to be the go-to source for economic research that creates opportunity for every member of the global workforce. Within the Economic Graph Research umbrella, we are interested in rigorously investigating economic and labor market phenomena including, but not limited to: human capital, career dynamics, and productivity growth.
AI: We include artificial intelligence in the core parts of the product and member experiences and are interested in a variety of AI problems, all with the goal of creating economic opportunity for every member of the global workforce. These interests include, but are not limited to, research into: large-scale machine learning, reinforcement learning, personalized machine learning, data mining (semi-structured and unstructured data, or time-series and spatial data), causal inference, and proposals related to the security and privacy applications of machine learning.
Full details about the submission process, selection criteria, and how we protect member privacy can all be found on the new Economic Graph Research website. Please join us in our endeavor to answer fundamental questions about data science and the world economy.