A Data-Driven Approach to Enhancing Gravity Models for Trip Demand Prediction

Houbing Song; Kamal Acharya; Liang Sun; Mehul Lad

REVIEW 3 major objections 5 minor 19 references

A Data-Driven Approach to Enhancing Gravity Models for Trip Demand Prediction

T0 review · 3 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Machine-learning trip-demand models fed with 27 county-level features consistently beat the same models given only the gravity model's classic inputs.

desk verdict The paper's featured improvement is an artifact of comparing an ML model with 27 features to the same ML model with 3 features; the gravity model equations are never actually fitted. read the letter →

arxiv 2506.01964 v1 pith:EM6KVUM7 submitted 2025-05-09 cs.LG

Kamal Acharya , Mehul Lad , Liang Sun , Houbing Song This is my paper · ORCID

classification cs.LG

keywords gravitymodeltripdemandpredictionmachinelearningneuralnetworksrandomforestgradientboostinghumanmobilityflowsSHAPfeatureimportance

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper argues that feeding trip-demand models a richer set of county-level features—land use, points of interest, roads, terminals, economic conditions, education, and population—makes them substantially more accurate than models that use only the gravity model's classic inputs of population, distance, and travel time. Using machine-learning regressors on data from Tennessee and New York counties, the authors report that a neural network in New York improves $R^2$ by 51.48%, cuts mean absolute error by 63.59%, and raises the Common Part of Commuters similarity measure by 44.32% relative to the baseline. The claim matters because transportation planners need reliable origin–destination trip forecasts for infrastructure, transit scheduling, and resource allocation, and the paper proposes a practical recipe: combine open geographic and socioeconomic datasets with off-the-shelf machine-learning algorithms. The broader thesis is that the rigid functional form of the gravity model can be effectively replaced by learned, non-linear mappings over many more variables.

What carries the argument

The object that carries the argument is a feature-augmented machine-learning gravity model: instead of fitting the closed-form gravity equation $T_{ij} = k P_i^{\lambda} P_j^{\alpha} / d_{ij}^{\beta}$, the paper trains regressors on the general form $T_{ij} = f(\mathbf{O}_i, \mathbf{D}_j, \mathbf{S}_{ij})$, where $\mathbf{O}_i$ and $\mathbf{D}_j$ are vectors of origin and destination attributes and $\mathbf{S}_{ij}$ captures separation. The concrete instantiation is a 27-feature dataset built by merging the origin-destination mobility-flow data, routing-derived distance and travel time, and county-level economic, education, and population data. This dataset replaces the sparse Dataset 1 (population, distance, travel time) and is fed into Random Forest, Gradient Boosting, and a five-layer neural network with ReLU activations, dropout, and Adam optimization. The contrast between the same algorithm on Dataset 1 versus Dataset 2 is what isolates the contribution of the extra features.

What would settle it

Fit the gravity model equation with calibrated constants and exponents on the same training data and evaluate it on the same held-out test set with the same three metrics; if a properly fitted gravity baseline matches or beats the Dataset-1 machine-learning model, then the headline improvements reflect the value of added features rather than a genuine advantage over the gravity model.

Watch

Extended reading notes

Core claim

The central discovery is that the addition of 27 features spanning eight categories—land use, points of interest, roads, transport terminals, building structures, economic factors, education, and population—to the inputs of standard machine-learning regressors yields consistently better trip-demand predictions than the same regressors trained only on the gravity model's three classic inputs (origin and destination population, distance, and travel time). In the paper's comparison, the 'traditional' gravity model is implemented as each machine-learning algorithm trained on Dataset 1, and the data-driven model is the same algorithm trained on Dataset 2 with all 27 features. Across Random Forest, Gradient Boosting, and especially a five-layer neural network, the data-driven models improve all three evaluation metrics for both Tennessee and New York; the largest gains appear in New York with the neural network, where $R^2$ rises from 0.6444 to 0.9762, mean absolute error falls from 0.0879 to 0.0320, and Common Part of Commuters rises from 0.6295 to 0.9085. The paper interprets this as evidence that non-linear learners can absorb diverse geographic, economic, and social data to overcome the gravity model's rigidity.

Load-bearing premise

The reported improvements assume that a machine-learning model trained only on population, distance, and travel time adequately represents the traditional gravity model, even though the paper never fits the gravity model's own equation as the baseline.

Editorial extensions

If this is right

In dense, data-rich urban settings, a neural-network-based data-driven model can raise $R^2$ from about 0.64 to about 0.98 for county-to-county trip flows, cutting mean absolute error by more than half.
For Random Forest, the added features produce only modest gains, meaning ensemble tree methods already capture much of the relevant structure from the gravity model's classic inputs.
Feature-attribution analysis indicates that travel time, distance, destination population, and destination education are consistently among the top predictors, so these variables are the ones worth collecting first.
Trip segmentation shows neural networks handle short and medium trips well in urban settings but degrade on long trips in sparse regions, while Random Forest stays accurate across trip lengths and weekdays/weekends.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

Inference: because the baseline 'traditional model' is a machine-learning regressor trained only on the gravity model's classic inputs rather than a fitted gravity equation, a head-to-head comparison with a calibrated gravity model could produce different improvement numbers than the 51.48% reported.
Inference: the larger gains in New York than Tennessee suggest the benefit of added features may grow with urban density and data richness, which could be tested by applying the same pipeline to additional states.
Inference: the paper's single-month mobility window leaves open the possibility that the reported improvements vary seasonally; repeating the analysis on multiple months or years would test the stability of the feature-augmented models.
Inference: the explicit two-dataset setup (Dataset 1 versus Dataset 2) offers a reusable benchmarking pattern for future trip-demand studies, independent of which regressor is used.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Two things to know. The experiments are a clean feature-ablation study for county-level trip flows in TN and NY: same ML models trained on 3 features versus 27 features. The framing, however, is wrong. The paper calls the 3-feature model the 'traditional gravity model' but never fits the gravity model equations (Eqs. 1-2). No k, lambda, alpha, or beta is estimated. So the headline 51% R2 improvement, 64% MAE reduction, 44% CPC increase are all relative to a feature-poor ML model, not to a gravity baseline. This is a load-bearing mismatch: the abstract's central claim is untested.

What the paper does well: it assembles a broad set of publicly available county features, runs three standard ML regressors, reports three metrics, and adds SHAP feature importance plus trip segmentation and weekday/weekend breakdowns. The metrics look computed correctly, and the feature importance lists are plausible. As a recipe for feature engineering in Tennessee and New York, it's usable.

Soft spots, in order. The baseline mislabeling is fatal to the paper's stated contribution. The 'improvement' numbers are only as strong as the weakest baseline, and a properly calibrated gravity model could well beat a 3-feature untuned ML regressor, especially on long-tail OD pairs. Second, the abstract cherry-picks the best result: the NY neural network. Random Forest and Gradient Boosting improvements are much smaller (0.4-5% R2 gains). Third, there are no error bars or repeated splits; from the description, it's a single train/test split, so the small gains could be noise. Fourth, they cite the deep gravity model (Ref. 14) but never compare against it or any classic gravity implementation, which would be the natural benchmark.

Who is this for? Transportation planners who want a quick template for county-level trip prediction with ML might get some value from the feature lists and model comparisons. Readers who know the ML-for-mobility literature will find nothing new: feature-ablation gains for ML regressors are already documented. The paper does not answer the question its title asks.

If this came to me for review, I'd reject. The authors could reframe it as a feature-ablation study and add a fitted gravity baseline; then it might be a modest empirical contribution. As it stands, the central claim is unsupported.

Referee Report

3 major / 5 minor

Summary. The paper proposes a data-driven extension of the gravity model for county-to-county trip demand prediction in Tennessee and New York. The authors assemble 27 features spanning land use, points of interest, roads, terminals, structures, economics, education, and population, and train three machine-learning models (random forest, deep neural network, and gradient boosting) on these features. They report improvements over what they call the "traditional" gravity model, with headline numbers in Table II such as a 51.48% R-squared improvement for the New York neural network. The methodology section states that Dataset 1, used for the traditional baseline, contains only population, distance, and travel time, while Dataset 2 adds the remaining features. The paper also reports SHAP-based feature importance and trip-distance/weekday-weekend segmentation analyses. The central claim is that machine-learning-enhanced models significantly outperform the traditional gravity model.

Significance. If the central claim were supported, the paper would offer a practical recipe for improving trip demand prediction with readily available geographical and socioeconomic data, and the CPC-based reliability analysis would be a useful addition. Some elements are genuinely valuable: the use of real mobility-flow data, the breadth of features considered, the comparison of three ML algorithms, and the attempt to interpret predictions with SHAP. However, the significance is sharply limited by the design of the baseline. The "traditional gravity model" is never fitted; it is replaced by the same ML algorithms trained on a reduced feature set. Consequently, the reported "improvements" are properly interpreted as a feature-ablation study, not as evidence that ML enhances the gravity model. The headline claim about outperforming the traditional model is therefore not established by the experiments as designed.

major comments (3)

[Section III.C / Table II] The central claim that machine-learning-enhanced models outperform the traditional gravity model is not tested. In Section III.B, Dataset 1 is defined as containing only population, distance, and travel time, and Section III.C states that the same ML algorithms (random forest, neural network, gradient boosting) are trained on this dataset. Table II labels these results as "Traditional." Nowhere in the paper are the parameters k, lambda, alpha, or beta of Eq. (2) estimated, and no log-linear, Poisson, or other fitted gravity-model predictions are evaluated. Thus the reported gains, including the 51.48% R-squared improvement in Table II, compare a 3-feature ML model with a 27-feature ML model. This can support a feature-ablation claim, but it does not support the abstract's and conclusion's claim that the data-driven approach outperforms the traditional gravity model. A properly calibrated gravity baseline could perform substantially differently, especially for sparse long-distance OD pairs, so the headline improvements could shrink or disappear under the correct comparison.
[Section III.B / Eqs. (1)-(2)] Even as a proxy for the gravity model, the "traditional" baseline is not faithful to the equations presented. Equation (2) specifies T_ij = k P_i^lambda P_j^alpha / d_ij^beta, which involves only origin population, destination population, and distance. Dataset 1, however, includes travel time as an additional input feature. The data-driven model is therefore compared not against the gravity model of Eq. (2) but against an ML model with an extra predictor beyond the gravity model's explanatory variables. This further weakens the interpretation of Table II as measuring enhancement of the gravity model.
[Abstract / Conclusion] The abstract and conclusion state that machine-learning models significantly outperform the traditional model and cite percentage improvements, but these statements are not supported by the experimental design described in Section III. The paper should either reframe its contribution as an ML feature-engineering/ablation study, or it must redo the evaluation with a genuinely fitted gravity baseline (including production-ready estimates of the distance-decay and population exponents). Without one of these changes, the paper's stated contribution is not validated.

minor comments (5)

[Section III.C] The evaluation protocol is incompletely specified: the paper does not state how the data are split into training and test sets, whether the split is stratified by state or by trip-length distribution, or whether hyperparameter tuning is nested inside cross-validation. Reporting standard deviations or confidence intervals across repeated splits would also strengthen the comparison.
[Table II] The column heading "Traditional" is misleading under the current implementation. If the paper is reframed as a feature-ablation study, the columns should be renamed accordingly, for example "Dataset 1 (3 features)" and "Dataset 2 (27 features)."
[Section II, Eq. (2)] Equation (2) is presented as the generalized gravity model but is never used in the experiments; the text should explicitly state that the ML baseline is not a fitted version of Eq. (2), or the equation should be replaced by a description of the actual baseline.
[Figures 2-3] The figures should include axis labels, units, and ideally error bars or confidence bands; currently the caption text describes trends that cannot be verified from the figure alone.
[Section V] The future-work sentence contains a grammatical error ("explore the incorporating real-time data"); it should read "explore incorporating real-time data."

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: the empirical ML comparisons do not reduce to their inputs by construction.

full rationale

I checked the derivation chain from Eqs. (1)-(3) through the Dataset 1/2 construction, Table II, and the reference list. No fitted parameter is renamed as a prediction, no quantity in the paper is defined in terms of the quantity it is supposed to predict, and there is no load-bearing self-citation or imported uniqueness theorem. The only questionable point is that the 'Traditional' baseline in Table II is implemented by training the same ML regressors on Dataset 1 (population, distance, travel time) instead of estimating the gravity equation of Eq. (2); this is a threat to construct validity and a serious correctness concern, because the reported 'improvements' may largely reflect the value of added features rather than an enhancement over the gravity model. However, a mislabeled or weak baseline is not circularity: the comparison is still an empirical test-set measurement, and the improvements do not follow from the equations by construction. Since the paper does not invoke any self-citation to force its conclusion, the circularity score is 0.

Assumptions & free parameters 8 free parameters · 4 assumptions · 0 invented entities

The central claim rests on two domain assumptions about data quality and feature relevance, plus the ad hoc assumption that a feature-poor ML model represents the gravity model. The ML hyperparameters are fitted to the data but are not the core of the claim. No theoretical constructs are invented.

free parameters (8)

Neural network learning rate = 0.00097
Tuned with Optuna to minimize validation MAE; reported in Section III.C.2.
Neural network dropout rate = 0.111
Tuned with Optuna; reported in Section III.C.2.
Neural network batch size = 64
Tuned with Optuna; reported in Section III.C.2.
Gradient boosting learning rate = 0.05
Selected by RandomizedSearchCV; reported in Section III.C.3.
Gradient boosting number of estimators = 500
Selected by RandomizedSearchCV; reported in Section III.C.3.
Gradient boosting max depth = 5
Selected by RandomizedSearchCV; reported in Section III.C.3.
Gradient boosting subsample = 0.9
Selected by RandomizedSearchCV; reported in Section III.C.3.
Random Forest hyperparameters = not reported
RandomizedSearchCV over 20 iterations with 5-fold CV was used, but the winning hyperparameter values are not given in the text.

assumptions (4)

domain assumption SafeGraph mobile phone mobility flows accurately represent county-level trip demand.
The dataset is used as ground truth without validation against travel surveys or other benchmarks. Section III.B.
domain assumption The 27 county features are appropriate and sufficient predictors of trip flows.
Features are chosen by availability, not by a mobility theory or prior causal model. Section III.B, Table I.
ad hoc to paper An ML model using only population, distance, and travel time is a faithful proxy for the traditional gravity model.
This assumption is needed for the 'Traditional' baseline in Table II to represent the gravity model, but the paper never fits the gravity model equation (Eq. 2). Section III.C.
domain assumption The train/test split is random and does not leak future information into features.
The split is not described, but supervised evaluation requires this. Section IV.

how reviews work

0 comments

Cite this review

Pith. "Pith review of A Data-Driven Approach to Enhancing Gravity Models for Trip Demand Prediction." pith.science (2026). https://pith.science/paper/EM6KVUM7

@misc{pith2026250601964,
  author       = {Pith},
  title        = {Pith review of: A Data-Driven Approach to Enhancing Gravity Models for Trip Demand Prediction},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/EM6KVUM7}},
  note         = {Machine review of arXiv:2506.01964}
}

read the original abstract

Accurate prediction of trips between zones is critical for transportation planning, as it supports resource allocation and infrastructure development across various modes of transport. Although the gravity model has been widely used due to its simplicity, it often inadequately represents the complex factors influencing modern travel behavior. This study introduces a data-driven approach to enhance the gravity model by integrating geographical, economic, social, and travel data from the counties in Tennessee and New York state. Using machine learning techniques, we extend the capabilities of the traditional model to handle more complex interactions between variables. Our experiments demonstrate that machine learning-enhanced models significantly outperform the traditional model. Our results show a 51.48% improvement in R-squared, indicating a substantial enhancement in the model's explanatory power. Also, a 63.59% reduction in Mean Absolute Error (MAE) reflects a significant increase in prediction accuracy. Furthermore, a 44.32% increase in Common Part of Commuters (CPC) demonstrates improved prediction reliability. These findings highlight the substantial benefits of integrating diverse datasets and advanced algorithms into transportation models. They provide urban planners and policymakers with more reliable forecasting and decision-making tools.

Figures

Figures reproduced from arXiv: 2506.01964 by the authors.

**Figure 1.** Research Process A. Study Area We selected two states, TN and NY, as the study area. TN has been chosen due to its diverse geographic and demographic composition, making it an ideal region for studying and modeling trip demand patterns. Similarly, NY has been selected as a complementary area of focus due to its highdensity urban environments and extensive transportation networks. Together, TN and NY offer distinc… view at source ↗

**Figure 2.** MAE Comparison of Models for NY and TN by Trip Distance [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

**Figure 3.** MAE Comparison of Models for NY and TN by Days [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

19 extracted references · 13 canonical work pages

[1]

Effects of built envi- ronment and weather on demands for transportation network company trips,

M. S. Hasnine, J. Hawkins, and K. N. Habib, “Effects of built envi- ronment and weather on demands for transportation network company trips,” Transportation Research Part A: Policy and Practice , vol. 150, pp. 171–185, 2021

work page 2021
[2]

The gravity model,

J. E. Anderson, “The gravity model,” Annu. Rev. Econ. , vol. 3, no. 1, pp. 133–160, 2011

work page 2011
[3]

Sanayei, A

R. Sanayei, A. Vafaeinejad, J. Karami, and H. Aghamohammadi Zan- jirabad, “A model development on gis-driven data to predict temporal daily collision through integrating discrete wavelet transform (dwt) and artificial neural network (ann) algorithms; case study: Tehran-qazvin freeway,”Geocarto International, vol. 37, no. 14, pp. 4141–4157, 2022

work page 2022
[4]

A statistical theory of spatial distribution models,

A. Wilson, “A statistical theory of spatial distribution models,” Trans- portation Research, vol. 1, no. 3, pp. 253–269, 1967. [Online]. Available: https://www.sciencedirect.com/science/article/pii/0041164767900354

arXiv 1967
[5]

Erlander and N

S. Erlander and N. F. Stewart, The gravity model in transportation analysis: theory and extensions . Vsp, 1990, vol. 3

work page 1990
[6]

A. S. Fotheringham and M. E. O’Kelly, Spatial interaction models: formulations and applications. Kluwer Academic Publishers Dordrecht, 1989, vol. 1

work page 1989
[7]

Gravity model in the korean highway,

W.-S. Jung, F. Wang, and H. E. Stanley, “Gravity model in the korean highway,” Europhysics Letters, vol. 81, no. 4, p. 48005, 2008

2008
[8]

Gravity and spatial interaction models,

K. E. Haynes and A. S. Fotheringham, “Gravity and spatial interaction models,” 2020

work page 2020

Show all 19 references

[9]

A gis toolkit for exploring geographies of household activity/travel behavior,

R. N. Buliung and P. S. Kanaroglou, “A gis toolkit for exploring geographies of household activity/travel behavior,” Journal of Transport Geography, vol. 14, no. 1, pp. 35–51, 2006

2006
[10]

Urban flow prediction from spatiotemporal data using machine learning: A survey,

P. Xie, T. Li, J. Liu, S. Du, X. Yang, and J. Zhang, “Urban flow prediction from spatiotemporal data using machine learning: A survey,” Information Fusion, vol. 59, pp. 1–12, 2020

2020
[11]

Ai-based neural network models for bus passenger demand forecasting using smart card data,

S. Liyanage, R. Abduljabbar, H. Dia, and P.-W. Tsai, “Ai-based neural network models for bus passenger demand forecasting using smart card data,” Journal of Urban Management, vol. 11, no. 3, pp. 365–380, 2022

2022
[12]

Examining the spatial- temporal relationship between urban built environment and taxi rider- ship: Results of a semi-parametric gwpr model,

C. Chen, T. Feng, C. Ding, B. Yu, and B. Yao, “Examining the spatial- temporal relationship between urban built environment and taxi rider- ship: Results of a semi-parametric gwpr model,” Journal of Transport Geography, vol. 96, p. 103172, 2021

2021
[13]

Travel time prediction using machine learning and weather impact on traffic conditions,

B. Deb, S. R. Khan, K. T. Hasan, A. H. Khan, and M. A. Alam, “Travel time prediction using machine learning and weather impact on traffic conditions,” in 2019 IEEE 5th International Conference for Convergence in Technology (I2CT). IEEE, 2019, pp. 1–8

2019
[14]

A deep gravity model for mobility flows generation,

F. Simini, G. Barlacchi, M. Luca, and L. Pappalardo, “A deep gravity model for mobility flows generation,” Nature communications, vol. 12, no. 1, p. 6576, 2021

2021
[15]

Mdlf: A multi-view-based deep learning framework for individual trip destination prediction in public transportation systems,

J. Zhao, L. Zhang, J. Ye, and C. Xu, “Mdlf: A multi-view-based deep learning framework for individual trip destination prediction in public transportation systems,” IEEE Transactions on Intelligent Transportation Systems, vol. 23, no. 8, pp. 13 316–13 329, 2021

2021
[16]

Using machine learning for direct demand modeling of ridesourcing services in chicago,

X. Yan, X. Liu, and X. Zhao, “Using machine learning for direct demand modeling of ridesourcing services in chicago,” Journal of Transport Geography, vol. 83, p. 102661, 2020

2020
[17]

Multiscale dynamic human mobility flow dataset in the u.s. during the covid-19 epidemic,

Y . Kang, S. Gao, Y . Liang, M. Li, and J. Kruse, “Multiscale dynamic human mobility flow dataset in the u.s. during the covid-19 epidemic,” Scientific Data, pp. 1–13, 2020

2020
[18]

Real-time routing with openstreetmap data,

D. Luxen and C. Vetter, “Real-time routing with openstreetmap data,” in Proceedings of the 19th ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems , ser. GIS ’11. New York, NY , USA: ACM, 2011, pp. 513–516. [Online]. Available: http://doi.acm...

2011
[19]

A unified approach to interpreting model predictions,

S. Lundberg, “A unified approach to interpreting model predictions,” arXiv preprint arXiv:1705.07874 , 2017

2017 arXiv

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.

[1] [1]

Effects of built envi- ronment and weather on demands for transportation network company trips,

M. S. Hasnine, J. Hawkins, and K. N. Habib, “Effects of built envi- ronment and weather on demands for transportation network company trips,” Transportation Research Part A: Policy and Practice , vol. 150, pp. 171–185, 2021

work page 2021

[2] [2]

The gravity model,

J. E. Anderson, “The gravity model,” Annu. Rev. Econ. , vol. 3, no. 1, pp. 133–160, 2011

work page 2011

[3] [3]

Sanayei, A

R. Sanayei, A. Vafaeinejad, J. Karami, and H. Aghamohammadi Zan- jirabad, “A model development on gis-driven data to predict temporal daily collision through integrating discrete wavelet transform (dwt) and artificial neural network (ann) algorithms; case study: Tehran-qazvin freeway,”Geocarto International, vol. 37, no. 14, pp. 4141–4157, 2022

work page 2022

[4] [4]

A statistical theory of spatial distribution models,

A. Wilson, “A statistical theory of spatial distribution models,” Trans- portation Research, vol. 1, no. 3, pp. 253–269, 1967. [Online]. Available: https://www.sciencedirect.com/science/article/pii/0041164767900354

arXiv 1967

[5] [5]

Erlander and N

S. Erlander and N. F. Stewart, The gravity model in transportation analysis: theory and extensions . Vsp, 1990, vol. 3

work page 1990

[6] [6]

A. S. Fotheringham and M. E. O’Kelly, Spatial interaction models: formulations and applications. Kluwer Academic Publishers Dordrecht, 1989, vol. 1

work page 1989

[7] [7]

Gravity model in the korean highway,

W.-S. Jung, F. Wang, and H. E. Stanley, “Gravity model in the korean highway,” Europhysics Letters, vol. 81, no. 4, p. 48005, 2008

2008

[8] [8]

Gravity and spatial interaction models,

K. E. Haynes and A. S. Fotheringham, “Gravity and spatial interaction models,” 2020

work page 2020