LID Framework: A new method for geospatial and exploratory data analysis of potential innovation deter-minants at the neighborhood level

The geography of innovation offers a framework to understand how territorial characteristics shape innovation, often via spatial and cognitive proximity. Empirical research has focused largely on national and regional scales, while urban and sub-regional geographies receive less attention. Local studies typically rely on limited indicators (e.g., firm-level data, patents, basic socioeconomic measures), with few offering a systematic framework integrating urban form, mobility, amenities, and human-capital proxies at the neighborhood scale. Our study investigates innovation at a finer spatial resolution, going beyond proprietary or static indicators. We develop the Local Innovation Determinants (LID) database and framework to identify key enabling factors across regions, combining traditional government data with publicly available data via APIs for a more granular understanding of spatial dynamics shaping innovation capacity. Using exploratory big and geospatial data analytics and random forest models, we examine neighborhoods in New York and Massachusetts across four dimensions: social factors, economic characteristics, land use and mobility, morphology, and environment. Results show that alternative data sources offer significant yet underexplored potential to enhance insights into innovation dynamics. City policymakers should consider neighborhood-specific determinants and characteristics when designing and implementing local innovation strategies.

💡 Research Summary

The paper addresses a notable gap in the geography of innovation literature by moving the unit of analysis from national or regional scales down to the neighborhood level. The authors develop the Local Innovation Determinants (LID) framework and accompanying database, which merges traditional government statistics (population, education, income, employment, etc.) with open‑source spatial data harvested via APIs such as OpenStreetMap and Google Maps. The spatial granularity is set at the U.S. ZIP‑code level, providing a practical proxy for neighborhoods while allowing for cross‑source integration.

Data collection focuses on two U.S. states—New York and Massachusetts—and captures 35 independent variables grouped into four conceptual dimensions: (a) social and human‑capital factors, (b) economic characteristics, (c) land‑use and mobility infrastructure, and (d) morphology and environmental quality. The dependent variables are two widely used proxies for innovative activity: the number of granted patents and the rate of start‑up formation, both measured with a four‑year lag (variables from 2012 linked to outcomes in 2016) to mitigate simultaneity bias.

Methodologically, the study employs random‑forest ensemble learning to assess variable importance and to model non‑linear relationships without the need for explicit specification of interactions. The results demonstrate that both conventional determinants (e.g., share of highly educated residents, R&D employment, median income) and alternative, “big‑data” indicators (walkability indices, park coverage, density of cafés and co‑working spaces, building age) rank highly in predicting innovation outcomes. Notably, the density of informal “third‑places” such as cafés and co‑working hubs emerges as a strong predictor, underscoring the role of everyday social interaction spaces in knowledge spillovers.

Comparative analysis between the two states reveals systematic differences: in New York, public‑transport accessibility carries more predictive weight, whereas in Massachusetts, green‑space availability is more influential. These findings illustrate how local urban form and lifestyle patterns modulate the innovation ecosystem.

The authors argue that the LID framework offers a scalable, reproducible pipeline for urban analytics: data acquisition, cleaning, feature engineering, and model training are automated, enabling other researchers or municipal agencies to replicate the approach in different contexts. Policy implications are explicit—city planners should design neighborhood‑specific interventions, such as increasing the supply of third‑places or improving walkability, to boost local innovative capacity.

Limitations are candidly discussed. ZIP codes are administrative units that may not perfectly align with functional neighborhoods, and the reliance on patents and start‑up formation as the sole innovation metrics excludes service‑oriented or digital innovations. Moreover, the observational design precludes definitive causal inference. Future work is suggested to incorporate richer innovation indicators (e.g., scientific publications, digital product launches), apply longitudinal causal models, and explore finer spatial delineations (e.g., census blocks or custom polygons).

In sum, the paper makes three key contributions: (1) it constructs a novel, open‑source‑enhanced database for neighborhood‑level innovation analysis; (2) it demonstrates that alternative spatial data substantially improve predictive performance for innovation outcomes; and (3) it provides actionable insights for policymakers seeking data‑driven, place‑based innovation strategies.