Graph-based representation for identifying individual travel activities with spatiotemporal trajectories and POI data

Abstract

Individual daily travel activities (e.g., work, eating) are identified with various machine learning models (e.g., Bayesian Network, Random Forest) for understanding people’s frequent travel purposes. However, labor-intensive engineering work is often required to extract effective features. Additionally, features and models are mostly calibrated for individual trajectories with regular daily travel routines and patterns, and therefore suffer from poor generalizability when applied to new trajectories with more irregular patterns. Meanwhile, most existing models cannot extract features to explicitly represent regular travel activity sequences. Therefore, this paper proposes a graph-based representation of spatiotemporal trajectories and point-of-interest (POI) data for travel activity type identification, defined as Gstp2Vec. Specifically, a weighted directed graph is constructed by connecting regular activity areas (i.e., zones) detected via clustering individual daily travel trajectories as graph nodes, with edges denoting trips between pairs of zones. Statistics of trajectories (e.g., visit frequency, activity duration) and POI distributions (e.g., percentage of restaurants) at each activity zone are encoded as node features. Next, trip frequency, average trip duration, and average trip distance are encoded as edge weights. Then a series of feedforward neural networks are trained to generate low-dimensional embeddings for activity nodes through sampling and aggregating spatiotemporal and POI features from their multihop neighborhoods. Activity type labels collected via travel surveys are used as ground truth for backpropagation. The experiment results with real-world GPS trajectories show that Gstp2Vec significantly reduces feature engineering efforts by automatically learning feature embeddings from raw trajectories with minimal prepossessing efforts. It not only enhances model generalizability to receive higher identification accuracy on test individual trajectories with diverse travel patterns, but also obtains better efficiency and robustness. In particular, our identification of the most common daily travel activities (e.g., Dwelling and Work) for people with diverse travel patterns outperforms state-of-the-art classification models.

POIs-based public preferences mapping on imbalanced supply-demand of recreation services can support sustainable coastal beach management

“…As expected, many studies have demonstrated that POIs data perform well in understanding travel activities and public preferences. Liu et al (2022) employed POIs distribution as a key Zoning of the supply and demand for beach recreation services in DPND. Li et al 10.3389/fmars.2024.1288112 node for identifying individual travel activities.…”

Where You Are Is What You Do: On Inferring Offline Activities From Location Data

“…Similarly in [28] graph convolutional neural networks (GCNs) [18] are used to classify GPS traces of 139 subjects into 5 offline activities. GCNs are also used recently in [26] to infer 8 offline activities in GPS diaries collected and labeled by 167 subjects.…”

“…It is well established that a user's location is indicative of the type of real-world, day-to-day activities, such as shopping and dining (Hereafter referred to as "offline activities") they perform. Research has shown that basic offline activities can be inferred from GPS traces using both conventional statistical methods [22][23][24] and machine learning algorithms [19,26,28,29]. The same has been demonstrated using mobile phone data collected city-wide at the base station level [25,30].…”

“…As with any type of location technology, opportunities exist for malicious targeting. For example, the same algorithm used in [26] to help patients with alcohol use disorder recover can be exploited to target users of a smartphone app most vulnerable to alcoholism. Such a risk is further amplified given the recent widespread availability of powerful machine learning algorithms [9] that require little to no knowledge of statistics and/or algorithm design to configure and employ.…”

Accurate Identification of Micro-scale Urban Functional Zones from the Perspective of Dialectical Materialism

“…The second issue concerns the different POI reclassification standards, where varying standards present challenges in conducting comparisons and analogies across research studies 42,43 . Inconsistent standards can lead to unreliable results due to the potential influence of personal preferences or assumptions rather than objective evidence in the identification process 44 .…”