Mining Points-of-Interest for Explaining Urban Phenomena: A Scalable Variational Inference Approach

Abstract

Points-of-interest (POIs; i.e., restaurants, bars, landmarks, and other entities) are common in web-mined data: they greatly explain the spatial distributions of urban phenomena. The conventional modeling approach relies upon feature engineering, yet it ignores the spatial structure among POIs. In order to overcome this shortcoming, the present paper proposes a novel spatial model for explaining spatial distributions based on web-mined POIs. Our key contributions are: (1) We present a rigorous yet highly interpretable formalization in order to model the influence of POIs on a given outcome variable. Specifically, we accommodate the spatial distributions of both the outcome and POIs. In our case, this modeled by the sum of latent Gaussian processes. (2) In contrast to previous literature, our model infers the influence of POIs without feature engineering, instead we model the influence of POIs via distance-weighted kernel functions with fully learnable parameterizations. (3) We propose a scalable learning algorithm based on sparse variational approximation. For this purpose, we derive a tailored evidence lower bound (ELBO) and, for appropriate likelihoods, we even show that an analytical expression can be obtained. This allows fast and accurate computation of the ELBO. Finally, the value of our approach for web mining is demonstrated in two real-world case studies. Our findings provide substantial improvements over state-of-the-art baselines with regard to both predictive and, in particular, explanatory performance. Altogether, this yields a novel spatial model for leveraging webmined POIs. Within the context of location-based social networks, it promises an extensive range of new insights and use cases. CCS CONCEPTS • Mathematics of computing → Variational methods; • Computing methodologies → Gaussian processes; • Information systems → Location based services; Web mining.

Housing prices and points of interest in three Polish cities

“…The ease of gathering data from different online sources (e.g., information on properties or sales, GIS data, coordinates, layers shared POI (e.g. parks, schools, restaurants, hotels, cinemas, monuments) is their cartographic mapping (geographic coordinates), which allows us to use these data in numerous research areas, such as Geographic Information Science (GIScience) (Gao et al, 2017;Wu et al, 2016), Urban Planning (Ganter et al, 2022;Naumzik et al, 2020), Socio-Economic (Dudás et al, 2017;Sun et al, 2022), Tourism and Marketing (Taylor et al, 2018;Yochum et al, 2020), Transportation (Jia et al, 2018) or Environmental Science (Dong et al, 2018).…”

Mining Points-of-Interest Data to Predict Urban Inequality: Evidence from Germany and France

“…The intuition behind density-based features is that some areas benefit from having many POIs of the same type in close vicinity (e.g., bars in nightlife districts). Here, a common choice is to count the number of POIs within a specific radius when making inferences in a city context (e.g., Hummler, Naumzik, and Feuerriegel 2022;Naumzik, Zoechbauer, and Feuerriegel 2020).…”

“…Prediction models: We also experimented with different neural networks; however, even with a high dropout rate, we experienced overfitting. We also considered the use of (Naumzik, Zoechbauer, and Feuerriegel 2020) but requires a spatial distribution as target variable, not data at neighborhood level, because of which this was inapplicable.…”