An integrated multivariate analysis—GIS framework for accurate microspatial energy demand forecasting

Adri Senen; Jasrul Jamani Jamian; Adri Senen; Jasrul Jamani Jamian

doi:10.3934/energy.2025046

AIMS Energy

2025, Volume 13, Issue 5: 1241-1272. doi: 10.3934/energy.2025046

Previous Article Next Article

Research article

An integrated multivariate analysis—GIS framework for accurate microspatial energy demand forecasting

Adri Senen ^{1,2
,
,},
Jasrul Jamani Jamian ¹

1.
Faculty of Electrical Engineering, Universiti Teknologi Malaysia, Johor, Malaysia
2.
Faculty of Electricity and Renewable Energy, Institut Teknologi PLN, Jakarta 1, Indonesia

Received: 26 June 2025 Revised: 06 September 2025 Accepted: 25 September 2025 Published: 10 October 2025

Current energy demand forecasting often cannot capture the complexity and diversity of electricity demand growth in smaller areas (microspatial). This challenge is exacerbated by variables affecting load growth that differ by location. Therefore, we proposed an integrated framework that blends multivariate analysis and geographic information systems (GIS) to generate more accurate and contextually relevant energy demand forecasts at the microspatial scale, employing a grid resolution of approximately 1.51 km × 1.51 km. Multivariate analysis was used to identify and select significant variables that represent the unique conditions of each area. The variables analyzed included demographic, geographic, economic, and sectoral electricity loads. The significant variables selected through this analysis were then used to form a predictive model of future energy demand growth. The results showed that four significant variables were identified from the initial ten variables for each cluster, with the model showing a high degree of accuracy (R²) of 0.9796, and a mean absolute percentage error (MAPE) was 3.36%. Moreover, GIS integration provided visualization and spatial analysis that strengthened the understanding of load distribution in various areas, thus supporting more effective network planning and decision-making. This approach showed significant potential in improving accuracy and spatial resolution in energy forecasting to support adaptive and sustainable electricity system management.
- energy demand forecasting,
- microspatial,
- cluster,
- multivariate,
- geographic information systems (GIS)
Citation: Adri Senen, Jasrul Jamani Jamian. An integrated multivariate analysis—GIS framework for accurate microspatial energy demand forecasting[J]. AIMS Energy, 2025, 13(5): 1241-1272. doi: 10.3934/energy.2025046

Related Papers:

Abstract

Current energy demand forecasting often cannot capture the complexity and diversity of electricity demand growth in smaller areas (microspatial). This challenge is exacerbated by variables affecting load growth that differ by location. Therefore, we proposed an integrated framework that blends multivariate analysis and geographic information systems (GIS) to generate more accurate and contextually relevant energy demand forecasts at the microspatial scale, employing a grid resolution of approximately 1.51 km × 1.51 km. Multivariate analysis was used to identify and select significant variables that represent the unique conditions of each area. The variables analyzed included demographic, geographic, economic, and sectoral electricity loads. The significant variables selected through this analysis were then used to form a predictive model of future energy demand growth. The results showed that four significant variables were identified from the initial ten variables for each cluster, with the model showing a high degree of accuracy (R²) of 0.9796, and a mean absolute percentage error (MAPE) was 3.36%. Moreover, GIS integration provided visualization and spatial analysis that strengthened the understanding of load distribution in various areas, thus supporting more effective network planning and decision-making. This approach showed significant potential in improving accuracy and spatial resolution in energy forecasting to support adaptive and sustainable electricity system management.

References

[1]	Evangelopoulos VA, Georgilakis PS (2022) Probabilistic spatial load forecasting for assessing the impact of electric load growth in power distribution networks. Electr Power Syst Res 207: 107847. https://doi.org/10.1016/j.epsr.2022.107847 doi: 10.1016/j.epsr.2022.107847
[2]	Thakare S, Bokde ND, Feijóo-Lorenzo AE (2023) Forecasting different dimensions of liquidity in the intraday electricity markets: A review. AIMS Energy 11: 918–959. https://doi.org/10.3934/ENERGY.2023044 doi: 10.3934/ENERGY.2023044
[3]	Jiang H, Dong Y, Dong Y, et al. (2024) Power load forecasting based on spatial-temporal fusion graph convolution network. Technol Forecast Soc Change 204: 123435. https://doi.org/10.1016/j.techfore.2024.123435 doi: 10.1016/j.techfore.2024.123435
[4]	Zhao Z, Zhang Y, Yang Y, et al. (2022) Load forecasting via grey model-least squares support vector machine model and spatial-temporal distribution of electric consumption intensity. Energy 255: 124468. https://doi.org/10.1016/j.energy.2022.124468 doi: 10.1016/j.energy.2022.124468
[5]	Zambrano-Asanza S, Morales RE, Montalvan JA, et al. (2023) Integrating artificial neural networks and cellular automata model for spatial-temporal load forecasting. Int J Electr Power Energy Syst 148: 108906. https://doi.org/10.1016/j.ijepes.2022.108906 doi: 10.1016/j.ijepes.2022.108906
[6]	Maswanganyi N, Ranganai E, Sigauke C (2019) Long-term peak electricity demand forecasting in South Africa: A quantile regression averaging approach. AIMS Energy 7: 857–882. https://doi.org/10.3934/energy.2019.6.857 doi: 10.3934/energy.2019.6.857
[7]	Senen A, Widyastuti C, Handayani O, et al. (2021) Development of micro-spatial electricity load forecasting methodology using multivariate analysis for dynamic area in Tangerang, Indonesia. Pertanika J Sci Tech 29: 2565–2578. https://doi.org/10.47836/pjst.29.4.18 doi: 10.47836/pjst.29.4.18
[8]	El khantach A, Hamlich M, Belbounaguia N (2019) Short-term load forecasting using machine learning and periodicity decomposition. AIMS Energy 7: 382–394. https://doi.org/10.3934/ENERGY.2019.3.382 doi: 10.3934/ENERGY.2019.3.382
[9]	Pérez-Chacón R, Asencio-Cortés G, Troncoso A, et al. (2024) Pattern sequence-based algorithm for multivariate big data time series forecasting: Application to electricity consumption. Future Gener Comput Syst 154: 397–412. https://doi.org/10.1016/j.future.2023.12.021 doi: 10.1016/j.future.2023.12.021
[10]	Tambunan HB, Hakam DF, Prahastono I, et al. (2020) The challenges and opportunities of renewable energy source (RES) penetration in Indonesia: Case study of Java-Bali power system. Energies 13: 1–22. https://doi.org/10.3390/en13225903 doi: 10.3390/en13225903
[11]	Trull O, García-Díaz JC, Troncoso A (2020) Initialization methods for multiple seasonal holt-winters forecasting models. Mathematics 8: 1–16. https://doi.org/10.3390/math8020268 doi: 10.3390/math8020268
[12]	Zhao P, Hu W, Cao D, et al. (2024) Enhancing multivariate, multi-step residential load forecasting with spatiotemporal graph attention-enabled transformer. Int J Electr Power Energy Syst 160: 110074. https://doi.org/10.1016/j.ijepes.2024.110074 doi: 10.1016/j.ijepes.2024.110074
[13]	Nie Y, Li P, Wang J, et al. (2024) A novel multivariate electrical price bi-forecasting system based on deep learning, a multi-input multi-output structure and an operator combination mechanism. Appl Energy 366: 123233. https://doi.org/10.1016/j.apenergy.2024.123233 doi: 10.1016/j.apenergy.2024.123233
[14]	Huang Y, Hasan N, Deng C, et al. (2022) Multivariate empirical mode decomposition based hybrid model for day-ahead peak load forecasting. Energy 239: 122245. https://doi.org/10.1016/j.energy.2021.122245 doi: 10.1016/j.energy.2021.122245
[15]	Arjmand A, Samizadeh R, Dehghani Saryazdi M (2020) Meta-learning in multivariate load demand forecasting with exogenous meta-features. Energy Effic 13: 871–887. https://doi.org/10.1007/s12053-020-09851-x doi: 10.1007/s12053-020-09851-x
[16]	Pearson K (1901) LIII. On lines and planes of closest fit to systems of points in space. The London, Edinburgh, and Dublin philosophical magazine and journal of science 2: 559–572. https://doi.org/10.1080/14786440109462720
[17]	Hu Y, Xia X, Fang J, et al. (2018) A multivariate regression load forecasting algorithm based on variable accuracy feedback. Energy Procedia 152: 1152–1157. https://doi.org/10.1016/j.egypro.2018.09.147 doi: 10.1016/j.egypro.2018.09.147
[18]	Efroymson MA (1970). Multiple regression analysis. In A. Ralston & H.S. Wilf (Eds.), Mathematical Methods for Digital Computers, 191–203. New York: Wiley.
[19]	Draper NR, Smith H (1981). Applied regression analysis (2nd ed.). New York: Wiley.
[20]	Brunsdon C, Fotheringham AS, Charlton ME (1996) Geographically weighted regression: A method for exploring spatial nonstationarity. Geogr Anal 28: 281–298. https://doi.org/10.1111/j.1538-4632.1996.tb00936.x doi: 10.1111/j.1538-4632.1996.tb00936.x
[21]	Hochreiter S, Schmidhuber J (1997) Long short-term memory. Neural Comput 9: 1735–1780. https://doi.org/10.1162/neco.1997.9.8.1735 doi: 10.1162/neco.1997.9.8.1735
[22]	Chen HF (2009) In silico log p prediction for a large data set with support vector machines, radial basis neural networks and multiple linear regression. Chem Biol Drug Des 74: 142–147. https://doi.org/10.1111/j.1747-0285.2009.00840.x doi: 10.1111/j.1747-0285.2009.00840.x
[23]	Zhang Q, Wang N, Cheng J, et al. (2020) A stepwise downscaling method for generating high-resolution land surface temperature from AMSR-E data. IEEE J Sel Top Appl Earth Obs Remote Sens 13: 5669–5681. https://doi.org/10.1109/JSTARS.2020.3022997 doi: 10.1109/JSTARS.2020.3022997
[24]	Żogała-Siudem B, Jaroszewicz S (2021) Fast stepwise regression based on multidimensional indexes. Inf Sci 549: 288–309. https://doi.org/10.1016/j.ins.2020.11.031 doi: 10.1016/j.ins.2020.11.031
[25]	Wang G, Sarkar A, Carbonetto P, et al. (2020) A simple new approach to variable selection in regression, with application to genetic fine mapping. J R Stat Soc Ser B Stat Methodol 82: 1273–1300. https://doi.org/10.1111/rssb.12388 doi: 10.1111/rssb.12388
[26]	Zeng S, Yan Z, Yang J (2022) Stepwise algorithm and new analytical model for estimating multi-parameter of energy piles from thermal response tests. Energy Build 256: 111775. https://doi.org/10.1016/j.enbuild.2021.111775 doi: 10.1016/j.enbuild.2021.111775
[27]	Subramanyam T, Donthi R, Satish Kumar V, et al. (2021) A new stepwise method for selection of input and output variables in data envelopment analysis. J Math Comput Sci 11: 703–715. https://doi.org/10.28919/jmcs/5205 doi: 10.28919/jmcs/5205
[28]	Zhuang J, Guo Z, Gong X, et al. (2023) Spatial load forecasting method based on GIS and generative adversarial network. 2023 8th Asia Conference on Power and Electrical Engineering (ACPEE), 1965–1969. https://doi.org/10.1109/ACPEE56931.2023.10135732
[29]	Katruksa S, Jiriwibhakorn S (2019) Electricity load forecasting based on a geographic information system. 2019 5th International Conference on Engineering, Applied Sciences and Technology (ICEAST), 1–4. https://doi.org/10.1109/ICEAST.2019.8802591
[30]	Messoudi L, Gouareh A, Settou B, et al. (2024) Modeling and forecasting energy consumption in Algerian residential buildings using a bottom-up GIS approach. Energy Build 317: 114370. https://doi.org/10.1016/j.enbuild.2024.114370 doi: 10.1016/j.enbuild.2024.114370
[31]	Zhao R, Zhan L, Yao M, et al. (2020) A geographically weighted regression model augmented by Geodetector analysis and principal component analysis for the spatial distribution of PM_2.5. Sustain Cities Soc 56: 102106. https://doi.org/10.1016/j.scs.2020.102106
[32]	Yasin H, Purhadi, Choiruddin A (2024) Spatial clustering based on geographically weighted multivariate generalized gamma regression. MethodsX 13: 102903. https://doi.org/10.1016/j.mex.2024.102903 doi: 10.1016/j.mex.2024.102903
[33]	Li W, Xu Q, Yi J, et al. (2021) Prediction and evaluation of forest fire in Yunnan of China based on geographically weighted logistic regression model[Preprint]. Res Square. https://doi.org/10.21203/rs.3.rs-997415/v1
[34]	Wu Y, Tang Z, Xiong S (2023) A unified geographically weighted regression model. Spat Stat 55: 100753. https://doi.org/10.1016/j.spasta.2023.100753 doi: 10.1016/j.spasta.2023.100753
[35]	Tephiruk N, Hongesombut K, Damrongkulkamjorn P, et al. (2023) Spatial electric load forecasting using a geographic information system: A case study of Khon Kaen, Thailand. 2023 IEEE 3rd International Conference on Sustainable Energy and Future Electric Transportation (SEFET), 1–7. https://doi.org/10.1109/SeFeT57834.2023.10245044
[36]	Liu Y, Li Y, Xu J, et al. (2021) Grid-based spatial load forecasting method based on POI information mining. 2021 IEEE 4th International Electrical and Energy Conference (CIEEC), 1–6. https://doi.org/10.1109/CIEEC50170.2021.9510563
[37]	Vieira DAG, Silva BE, Menezes TV, et al. (2020) Large scale spatial electric load forecasting framework based on spatial convolution. Int J Electr Power Energy Syst 117: 105582. https://doi.org/10.1016/j.ijepes.2019.105582 doi: 10.1016/j.ijepes.2019.105582
[38]	Si C, Xu S, Wan C, et al. (2021) Electric load clustering in smart grid: Methodologies, applications, and future trends. J Mod Power Syst Clean Energy 9: 237–252. https://doi.org/10.35833/MPCE.2020.000472 doi: 10.35833/MPCE.2020.000472
[39]	Senen A, Putri TWO, Jamian JJ, et al. (2023) Fuzzy C-means clustering based on micro-spatial analysis for electricity load profile characterization. Indones J Electr Eng Comput Sci 30: 33–45. https://doi.org/10.11591/ijeecs.v30.i1.pp33-45 doi: 10.11591/ijeecs.v30.i1.pp33-45
[40]	Bezdek JC (1981) Pattern recognition with fuzzy objective function algorithms. New York: Springer. https://doi.org/10.1007/978-1-4757-0450-1
[41]	Rousseeuw PJ (1987) Silhouettes: A graphical aid to the interpretation and validation of cluster analysis. J Comput Appl Math 20: 53–65. https://doi.org/10.1016/0377-0427(87)90125-7 doi: 10.1016/0377-0427(87)90125-7
[42]	Aswanuwath L, Pannakkong W, Buddhakulsomsiri J, et al. (2023) A hybrid model of VMD-EMD-FFT, similar days selection method, stepwise regression, and artificial neural network for daily electricity peak load forecasting. Energies 16: 1860. https://doi.org/10.3390/en16041860 doi: 10.3390/en16041860
[43]	Sethi R, Kleissl J (2020) Comparison of short-term load forecasting techniques. 2020 IEEE Conference on Technologies for Sustainability (SusTech), 1–6. https://doi.org/10.1109/SusTech47890.2020.9150490
[44]	Balzanella A, Verde R, Tenorio de Carvalho FA (2025) A cluster-wise regression method for distribution-valued data. Knowl-Based Syst 326: 113989. https://doi.org/10.1016/j.knosys.2025.113989 doi: 10.1016/j.knosys.2025.113989
[45]	Ingdal M, Johnsen R, Harrington DA (2019) The Akaike information criterion in weighted regression of immittance data. Electrochim Acta 317: 648–653. https://doi.org/10.1016/j.electacta.2019.06.030 doi: 10.1016/j.electacta.2019.06.030
[46]	Jing X, Qinggui C, Haiqin Q, et al. (2020) Assessment of aero-engine service reliability based on Akaike information criterion. 2020 5th International Conference on Mechanical, Control and Computer Engineering (ICMCCE), 636–639. https://doi.org/10.1109/ICMCCE51767.2020.00141
[47]	Portet S (2020) A primer on model selection using the Akaike Information Criterion. Infect Dis Model 5: 111–128. https://doi.org/10.1016/j.idm.2019.12.010 doi: 10.1016/j.idm.2019.12.010
[48]	Wu Y, Tang Z, Xiong S (2023) A unified geographically weighted regression model. Spat Stat 55: 1–25. https://doi.org/10.1016/j.spasta.2023.100753 doi: 10.1016/j.spasta.2023.100753
[49]	Chen F, Leung Y, Wang Q, et al. (2024) Spatial non-stationarity test of regression relationships in the multiscale geographically weighted regression model. Spat Stat 62: 100846. https://doi.org/10.1016/j.spasta.2024.100846 doi: 10.1016/j.spasta.2024.100846

Reader Comments

Your name:*

Email:*
© 2025 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)