09-20, 11:50–12:25 (Europe/Amsterdam), Van Gogh
We construct and analyze a time-varying worldwide network of professional relationships among startups to predict long-term economic performance using network centrality measures.
The presentation will provide an overview of the world wide startup network construction from CrunchBase data using Networkx. We will be modeling employee flow and knowledge transfer as links between startups. Using a network centrality we will be able to rank early stage startups (pre-seeded) and evaluate how their ranked position is correlated with their success ahead of time. Finally we will touch on implications for entrepreneurs, investors, and policymakers
By drawing on large-scale online data, we construct a time-varying worldwide network modeling the professional relationships and employee transitions among startup companies. The nodes represent companies, while links indicate the flow of employees and associated transfer of knowledge across firms.
We investigate whether the position and connectivity patterns of a startup within this network hold predictive power for the company's long-term economic performance and likelihood of becoming successful. Leveraging network centrality measures like PageRank, we rank all startups of the world wide startup network. Our analysis finds that the startup network provides valuable signals, enabling predictions that sometimes double the performance of traditional venture capital screening processes. The results support theories that a start-up's position within its ecosystem is highly relevant for future success.
This talk will cover the network construction methodology, predictive modeling approach, key findings, and implications for entrepreneurs seeking to optimize their start-up's ecosystem positioning. We also discuss how the analysis can aid venture capitalists and policymakers in conducting more objective assessments and targeted interventions within innovation ecosystems.
The target audience includes data scientists, network scientists, investors, entrepreneurs, and anyone interested in empirical studies combining network science and machine learning. The talk will be a mix of technical modeling details and higher-level insights. Basic prior knowledge of networks/graphs and machine learning concepts is recommended to fully appreciate the technical depths and lots of curiosity for using a new approach in the realm of Venture Capital funds!
I'm a data-driven problem solver with diverse interests spanning physics, venture capital, machine learning, and DIY robotics. My PhD research focused on investigating the evolution and success factors of complex social systems, particularly in the startup ecosystem.
Currently, I lead the data science team for consumer and affluent customers at ABN AMRO. By applying advanced data science techniques to challenges like churn prediction, propensity modeling, and recommender systems, I'm helping shape the bank of tomorrow.
My work, which I'll be presenting at PyData Amsterdam, has garnered attention in Wired UK and was published in Nature Scientific Reports. It explores the emergence of large-scale order in complex systems, offering insights applicable across various domains. Links below:
- https://www.wired.com/story/how-to-grow-startup/
- https://www.nature.com/articles/s41598-019-57209-w