Smart Out-of-Home Advertising Using Artificial Intelligence and GIS Data

Nader Nader, Rafael Alexandrou, Iasonas Iasonos, Andreas Pamboris, Harris Papadopoulos, Andreas KonstantinidisAAAI 2022 · DOI

Introduction

Out-of-Home (OOH) advertising, encompassing billboards and posters, is a multi-billion dollar industry projected to reach $33 billion by 2026. Despite its scale, selecting the optimal billboard locations has historically relied on "expert judgment" rather than quantitative data. This subjectivity often leads to arbitrary pricing and sub-optimal campaign performance. The Smart Out-of-Home Advertising Platform (SOAP) addresses these challenges by leveraging Geographic Information Systems (GIS) and machine learning to transform billboard selection into a data-driven science.

Quantifying Visibility with GIS and LiDAR

A core innovation of SOAP is its approach to visibility. Rather than simply noting proximity to a road, the platform uses airborne laser scanning (LiDAR) data to perform high-precision visibility analysis. By calculating "pixel counts" within specific radii, the system accounts for physical obstructions and the actual line-of-sight available to consumers. This allows the researchers to quantify visibility much more accurately than traditional methods.

Visibility analysis using LiDAR data to calculate line-of-sight and pixel counts around specific billboard sites.

Multi-Objective Optimization and the Pareto Front

Advertisers typically face conflicting goals: maximizing geographic coverage and audience similarity while minimizing costs. SOAP treats this as a Multi-Objective Optimization (MOO) problem. Using state-of-the-art algorithms like NSGA-II and MOEA/D, the platform generates a "Pareto Front"—a set of optimal solutions where improving one objective (e.g., coverage) necessitates a trade-off in another (e.g., cost).

The SOAP interface showing the Pareto Front trade-off between coverage and cost (left) and the resulting geographic distribution of billboards (right).

This approach empowers decision-makers by visualizing the cost-benefit ratio of different campaign configurations. Users can interactively explore these trade-offs, seeing how specific budget allocations translate to real-world geographic spread and demographic reach across the study area in Cyprus.

Data-Informed Repricing and Similarity

To move away from arbitrary industry pricing, SOAP employs Multiple Linear Regression to derive a data-informed pricing model. This model justifies costs based on tangible billboard characteristics, such as surrounding points of interest and traffic patterns. Additionally, the platform uses cosine similarity to match billboard profiles (location, surrounding demographics) with specific customer profiles gathered via questionnaires, ensuring campaigns reach the most relevant audience.

Conclusion and Future Work

The SOAP platform demonstrates the power of integrating disparate data sources—GIS, demographics, and LiDAR—into a unified AI framework. While currently focused on the Cyprus market, the methodology provides a scalable blueprint for modernizing the OOH advertising industry. Future work will look toward integrating real-time traffic data and closing the loop on ROI through footfall conversion tracking, further refining the "Visibility" and "Coverage" metrics for dynamic urban environments.

Summary updated

BibTeX
@inproceedings{nader2022smart,
  author = {Nader Nader and Rafael Alexandrou and Iasonas Iasonos and Andreas Pamboris and Harris Papadopoulos and Andreas Konstantinidis},
  title = {{Smart Out-of-Home Advertising Using Artificial Intelligence and GIS Data}},
  booktitle = {AAAI Conference on Artificial Intelligence},
  year = {2022},
  doi = {10.1609/aaai.v36i11.21729},
}
Figure