SIECCI — Predictive Modeling and Causality Analysis of Energy Consumption and Climate Change in Indonesia

00 - AboutTentang

About the ProjectTentang Proyek

Final project of D4 Applied Data Science, Politeknik Elektronika Negeri Surabaya — integrating predictive modeling, causality analysis, and an interactive dashboard in one unified SIECCI platform. Proyek akhir D4 Sains Data Terapan, Politeknik Elektronika Negeri Surabaya — mengintegrasikan pemodelan prediktif, analisis kausalitas, dan dashboard interaktif dalam satu platform SIECCI yang terpadu.

Student / AuthorMahasiswa / Penulis

Maulana Ihsan Athallah

NRP. 3322600012

Responsible for the entire research process: data collection and preprocessing, implementation of BVAR, Seq2Seq LSTM, and Holt Method models, Granger and Transfer Entropy causality analysis, and development of the SIECCI analytics dashboard using Tableau Public.Bertanggung jawab atas keseluruhan proses penelitian: pengumpulan dan pra-pemrosesan data, implementasi model BVAR, Seq2Seq LSTM, dan Holt Method, analisis kausalitas Granger dan Transfer Entropy, serta pengembangan dashboard analitik SIECCI menggunakan Tableau Public.

Supervisor IDosen Pembimbing I

Ronny Susetyoko, S.Si., M.Si

NIP. 197112111995011001

Guidance on multivariate statistics, time series modeling, stationarity testing, and quantitative analysis validation.Bimbingan metodologis pada aspek statistika multivariat, pemodelan deret waktu, uji stasioneritas, dan validasi analisis kuantitatif penelitian.

Supervisor IIDosen Pembimbing II

Alfi Fadliana, S.Si., M.Stat

NIP. 199211252022032018

Guidance on machine learning, deep learning modeling, hyperparameter optimization, and academic quality review of the research.Bimbingan pada aspek pemodelan machine learning, deep learning, optimasi hyperparameter, dan review kualitas akademik penelitian secara menyeluruh.

Project InformationInformasi Proyek

TitleJudulPredictive Modeling and Causality Analysis of Energy Consumption and Climate Change in IndonesiaPemodelan Prediktif dan Analisis Kausalitas pada Konsumsi Energi dan Perubahan Iklim di Indonesia

Study ProgramProgram StudiD4 Applied Data Science

DepartmentJurusanInformatics & Computer Engineering

InstitutionInstitusiPoliteknik Elektronika Negeri Surabaya (PENS)

YearTahun2026

Main MethodsMetode UtamaBVAR, Seq2Seq LSTM, Holt Method (DES)

CausalityKausalitasGranger Causality Test, Transfer Entropy

VisualizationVisualisasiTableau Public — SIECCI Dashboard

Data SourcesIEA, Our World in Data, IMF Climate Data

Data PeriodPeriode Data1990 – 2023 (34 yearstahun, annual)

About the Name SIECCITentang Nama SIECCI

SIECCI stands for System for Indonesian Energy Consumption & Climate Change Intelligence — a platform that not only presents data but also provides analytical intelligence to understand the complex interactions between energy consumption and climate change in Indonesia. It serves as a strategic, evidence-based tool for policymakers, researchers, and the general public in supporting SDGs 7 and SDGs 13 toward a cleaner and more sustainable Indonesia. SIECCI adalah singkatan dari System for Indonesian Energy Consumption & Climate Change Intelligence — sebuah platform yang tidak hanya menyajikan data, tetapi juga memberikan kecerdasan analitik untuk memahami interaksi kompleks antara konsumsi energi dan perubahan iklim di Indonesia. Platform ini hadir sebagai alat bantu strategis berbasis bukti bagi pengambil kebijakan, peneliti, dan masyarakat umum dalam mendukung pencapaian SDGs 7 dan SDGs 13.

Research DetailsDetail Penelitian

Click any section below to expandKlik salah satu bagian di bawah untuk membuka

01 · BackgroundLatar Belakang ▾ 02 · Research Questions & ObjectivesRumusan & Tujuan ▾ 03 · DatasetData ▾ 04 · MethodologyMetode ▾ 05 · FindingsTemuan ▾

01 — IntroductionPendahuluan

BackgroundLatar Belakang

In the era of Industry 5.0, Indonesia's energy consumption continues to grow yet remains dominated by fossil fuels, while climate change phenomena become increasingly evident. A data-driven approach is urgently needed to project and capture the complex relationships between these dimensions. Dalam era revolusi industri 5.0, konsumsi energi Indonesia terus meningkat signifikan namun masih didominasi energi fosil, sementara fenomena perubahan iklim semakin nyata. Diperlukan pendekatan berbasis data yang mampu memproyeksikan dan menangkap hubungan kompleks antar dimensi ini.

Energy is a fundamental component of human life and national development. Nearly all activities — from individual to institutional scale — require various forms of energy, from electricity and liquid fuels to gas. Macroeconomically, a nation's economic growth can be gauged by its availability and utilization of energy across industry, commercial, and residential sectors. Energi merupakan komponen utama dalam aktivitas kehidupan dan pembangunan suatu negara. Hampir seluruh aktivitas manusia, baik dalam skala individu maupun kelembagaan membutuhkan berbagai macam energi — mulai dari listrik, bahan bakar cair, hingga gas. Secara makro, pertumbuhan ekonomi suatu negara dapat dilihat dari ketersediaan dan pemanfaatan energi di berbagai sektor antara lain industri, komersial, dan rumah tangga.

Indonesia's energy consumption since 2010 has shown a fluctuating trend — rising from 1,738,335 TWh in 2010 to 2,299,630 TWh in 2019, temporarily declining due to the pandemic, then surging again post-2021 to reach 2,807,733 TWh in 2023 — still dominated by fossil fuels. Konsumsi energi Indonesia sejak 2010 menunjukkan tren fluktuatif — dari 1.738.335 TWh pada 2010 meningkat menjadi 2.299.630 TWh pada 2019, sempat menurun akibat pandemi, lalu kembali naik signifikan pasca-2021 hingga mencapai 2.807.733 TWh pada 2023. Namun, konsumsi ini masih didominasi energi fosil yang belum ramah lingkungan.

BMKG data shows an average temperature increase of 1.02 °C over 1981–2024 (0.23 °C per decade). Indonesia's CO₂ emissions from fossil fuels rose 13.14% in 2022 to 692 million tonnes, while global CO₂ emissions reached a record 36.8 gigatonnes. Data BMKG menunjukkan kenaikan suhu rata-rata sebesar 1,02 °C selama 1981–2024 atau setara 0,23 °C per dekade. Emisi CO₂ dari bahan bakar fosil di Indonesia naik 13,14% pada 2022 menjadi 692 juta ton, sementara emisi CO₂ global mencapai 36,8 gigaton — rekor tertinggi sepanjang sejarah.

As part of its SDGs commitment, Indonesia has adopted SDGs 7 (affordable and clean energy) and SDGs 13 (climate action). This is where data-driven predictive modeling and causality analysis become essential for generating adaptive, evidence-based policies. Sebagai komitmen terhadap SDGs, Indonesia telah mengadopsi SDGs 7 (energi bersih dan terjangkau) dan SDGs 13 (penanganan perubahan iklim). Di sinilah pendekatan prediktif berbasis data dan analisis kausalitas dibutuhkan untuk menghasilkan evidence-based policy yang lebih adaptif dan responsif.

2,807,733 TWh

Total Indonesia Energy Consumption (2023) — up from 1,738,335 TWh in 2010Total Konsumsi Energi Indonesia (2023) — naik dari 1.738.335 TWh pada 2010

Source: Our World in Data / IEA

+1.02 °C

Average surface temperature increase in Indonesia 1981–2024 (0.23 °C/decade)Kenaikan suhu permukaan rata-rata Indonesia selama 1981–2024 (0,23 °C/dekade)

Source: BMKG

692 Mt CO₂

CO₂ emissions from fossil fuels in Indonesia (2022) — up 13.14%Emisi CO₂ dari bahan bakar fosil Indonesia (2022) — naik 13,14%

Source: Worldometers

36.8 Gt

Global CO₂ emissions (2022) — all-time record, +0.5 Gt from 2021Emisi CO₂ global (2022) — rekor tertinggi sepanjang sejarah, naik 0,5 Gt dari 2021

Source: IEA, 2023

02 — IntroductionPendahuluan

Research Questions & ObjectivesRumusan Masalah & Tujuan

This study formulates three core research questions addressed through an integrated analytical approach — from predictive modeling and causality analysis to the development of an interactive visualization platform. Penelitian ini merumuskan tiga permasalahan inti yang dijawab melalui pendekatan analitik terintegrasi — dari pemodelan prediktif dan analisis kausalitas hingga pengembangan platform visualisasi interaktif berbasis Tableau.

Research QuestionsRumusan Masalah

What are the in-depth growth patterns of energy consumption in Indonesia, particularly across the industrial, residential, and commercial sectors?Bagaimana pola pertumbuhan konsumsi energi di Indonesia secara mendalam, khususnya untuk sektor industri, rumah tangga, dan komersial?

What is the quantitative relationship between energy consumption and climate change in Indonesia, including the direction and intensity of influence between them?Bagaimana keterkaitan kuantitatif antara konsumsi energi dan perubahan iklim di Indonesia, termasuk arah dan intensitas pengaruh di antara keduanya?

How can an interactive analytics platform be built that presents the energy-climate linkage in an integrated manner, specifically supporting SDGs 7 and SDGs 13?Bagaimana membangun platform analitik interaktif yang menyajikan keterkaitan energi-iklim secara terpadu untuk mendukung SDGs 7 dan SDGs 13?

Research ObjectivesTujuan Penelitian

Perform predictive modeling of energy consumption in industrial, household, and commercial sectors, as well as climate change indicators, using Bayesian Vector AutoRegression (BVAR), Sequence-to-Sequence (Seq2Seq), and Holt Method.Melakukan pemodelan prediktif terhadap konsumsi energi sektor industri, rumah tangga, dan komersial serta indikator perubahan iklim menggunakan metode Bayesian Vector AutoRegression (BVAR), Sequence-to-Sequence (Seq2Seq), dan Holt.

Analyze the causal relationships between energy consumption and climate change using Granger Causality Test and Transfer Entropy.Menganalisis hubungan kausalitas antara konsumsi energi terhadap perubahan iklim menggunakan Uji Kausalitas Granger dan Transfer Entropy.

Develop an interactive analytics dashboard using Tableau as a tool for monitoring energy consumption and climate change in Indonesia.Mengembangkan dashboard analitik menggunakan Tableau sebagai alat bantu monitoring konsumsi energi dan perubahan iklim di Indonesia.

Research BenefitsManfaat Penelitian

For the AuthorBagi Penulis

Deepen understanding of predictive and causality methods for energy and environmental data analysis.Meningkatkan pemahaman dalam penerapan metode prediktif dan kausalitas untuk analisis data energi dan lingkungan.

Develop skills in building comprehensive data-driven analytics information systems.Mengembangkan keterampilan membangun sistem informasi analitik berbasis data secara komprehensif.

For the GovernmentBagi Pemerintah

A data-driven tool for ESDM, KLHK, and Bappenas agencies to formulate, monitor, and evaluate SDGs 7 & 13 policies.Alat bantu berbasis data bagi instansi ESDM, KLHK, dan Bappenas dalam merumuskan, memantau, dan mengevaluasi kebijakan SDGs 7 & 13.

Supports more targeted decision-making through informative data visualization and accurate predictions.Mendukung pengambilan keputusan yang lebih tepat sasaran melalui visualisasi data yang informatif dan prediksi yang akurat.

For the PublicBagi Masyarakat Umum

Provides integrated energy consumption and climate change information transparently and accessibly to the public.Tersedia informasi konsumsi energi dan perubahan iklim terintegrasi secara transparan dan mudah diakses oleh publik.

Raises public literacy on energy efficiency and climate change awareness in Indonesia.Meningkatkan literasi masyarakat mengenai efisiensi energi dan kesadaran terhadap isu perubahan iklim di Indonesia.

03 — Dataset

Research Dataset StructureStruktur Data Penelitian

Annual time-series data (1990–2023) split into two groups: energy consumption per sector (13 variables) and climate change indicators (3 variables), sourced from IEA, Our World in Data, and IMF Climate Change Data. Data bersifat deret waktu tahunan (1990–2023) terbagi dalam dua kelompok: data konsumsi energi per sektor (13 variabel) dan data indikator perubahan iklim (3 variabel), bersumber dari IEA, Our World in Data, dan IMF Climate Change Data.

Energy Consumption — 13 variablesKonsumsi Energi — 13 variabel Climate Change — 3 variablesPerubahan Iklim — 3 variabel Period: 1990–2023 (34 years)Periode: 1990–2023 (34 tahun)

CodeVariableDescriptionDeskripsiSourceSumberUnit

Ind_Coal

Coal & coal product consumption — industrial sectorKonsumsi batubara & produk batubara sektor industri

IEA (sumber utama)

Ind_Natural_gas

Natural gas consumption — industrial sectorKonsumsi gas alam sektor industri

IEA (sumber utama)

Ind_Oil

Oil & oil products consumption — industrial sectorKonsumsi minyak & produk minyak sektor industri

IEA (sumber utama)

Ind_Biofuels

Biofuels & waste consumption — industrial sectorKonsumsi biofuel & limbah sektor industri

IEA (sumber utama)

Ind_Electricity

Electricity consumption — industrial sectorKonsumsi listrik sektor industri

IEA (sumber utama)

HH_OilRT_Oil

HH_Oil — oil product consumption, residential sectorKonsumsi minyak produk sektor rumah tangga (RT_Oil)

IEA (sumber utama)

HH_Natural_gasRT_Natural_gas

HH_Natural_gas — natural gas consumption, residential sectorKonsumsi gas alam sektor rumah tangga (RT_Natural_gas)

IEA (sumber utama)

HH_BiofuelsRT_Biofuels

HH_Biofuels — biofuels & waste consumption, residential sectorKonsumsi biofuel & limbah sektor rumah tangga (RT_Biofuels)

IEA (sumber utama)

HH_ElectricityRT_Electricity

HH_Electricity — electricity consumption, residential sectorKonsumsi listrik sektor rumah tangga (RT_Electricity)

IEA (sumber utama)

x10

Com_OilKom_Oil

Oil consumption — commercial sectorKonsumsi minyak sektor komersial

IEA (sumber utama)

x11

Com_BiofuelsKom_Biofuels

Biofuels & waste — commercial sectorKonsumsi biofuel & limbah sektor komersial

IEA (sumber utama)

x12

Com_Natural_gasKom_Natural_gas

Natural gas — commercial sectorKonsumsi gas alam sektor komersial

IEA (sumber utama)

x13

Com_ElectricityKom_Electricity

Electricity — commercial sectorKonsumsi listrik sektor komersial

IEA (sumber utama)

Emission_CO₂

Annual CO₂ emissions — IndonesiaEmisi karbon dioksida tahunan Indonesia

Our World in Data

Tonnes

Emission_GRK

Greenhouse Gas emissions (CO₂ equivalents)Emisi gas rumah kaca (CO₂ equivalents)

Our World in Data

Tonnes

Temperature

Annual surface temperature anomaly — IndonesiaAnomali suhu permukaan tahunan Indonesia

IMF Climate Change Data

°C

04 — MethodologyMetodologi

Research MethodologyMetode Penelitian

A combination of multivariate predictive and causality analysis approaches to generate a comprehensive picture of the dynamic relationship between energy consumption and climate change in Indonesia. Kombinasi pendekatan prediktif multivariat dan analisis kausalitas untuk menghasilkan gambaran komprehensif dinamika hubungan konsumsi energi dan perubahan iklim di Indonesia beserta proyeksi masa depannya.

Forecasting · MultivariatePeramalan · Multivariat

Bayesian Vector AutoRegression

BVAR model with Minnesota Prior and Gibbs Sampling (12,000 draws, 7,000 burn-in). Captures dynamic inter-variable relationships with high stability even with limited observations. Best for Temperature and Emission_GRK forecasting.Model BVAR dengan Minnesota Prior dan Gibbs Sampling (12.000 draws, 7.000 burn-in). Mampu menangkap hubungan dinamis antar variabel dengan stabilitas tinggi meski observasi terbatas. Unggul untuk peramalan Temperature dan Emission_GRK.

BVAR — Minnesota Prior

Forecasting · Deep LearningPeramalan · Deep Learning

Sequence-to-Sequence LSTM

Encoder-decoder LSTM architecture for non-linear temporal patterns. Hyperparameters optimized using Optuna (128 units, tanh, lr=0.001, batch=8). Best model for Emission_CO₂ forecasting with MAPE 10.57%.Arsitektur encoder-decoder berbasis LSTM untuk pola temporal non-linear. Hyperparameter dioptimasi menggunakan Optuna (128 units, tanh, lr=0.001, batch=8). Terbukti paling optimal untuk peramalan Emission_CO₂ dengan MAPE 10,57%.

Seq2Seq + Optuna Tuning

Forecasting · BenchmarkPeramalan · Benchmark

Double Exponential Smoothing

Holt Method capturing level and trend components adaptively. Parameters optimized via Grid Search (α, β: 0.01–0.99). Consistently best model (lowest MAE) across all 13 annual energy consumption variables.Metode Holt yang menangkap komponen level dan tren secara adaptif. Parameter dioptimasi via Grid Search (α, β: 0.01–0.99). Konsisten terbaik (MAE terendah) pada seluruh 13 variabel konsumsi energi tahunan.

Holt Method + Grid Search

Causality · LinearKausalitas · Linier

Granger Causality Test

Statistical hypothesis test for the direction of linear causal influence between time-series variables. Tested at lags 1–5 with significance level α=10%. Identifies more causal pairs than Transfer Entropy.Pengujian hipotesis statistik untuk arah pengaruh kausalitas linier antar variabel deret waktu. Diuji pada lag 1–5 dengan tingkat signifikansi α=10%. Mengidentifikasi lebih banyak hubungan kausal dibanding Transfer Entropy.

F-Statistic, p-value ≤ 0.10

Causality · Non-linearKausalitas · Non-linier

Transfer Entropy

Information-theoretic method based on Shannon Entropy to detect non-linear information flow between variables. Reveals relationships undetected by Granger — especially temperature's influence on natural gas consumption.Metode information-theoretic berbasis Shannon Entropy untuk mendeteksi aliran informasi non-linear antar variabel. Mengungkap hubungan yang tidak terdeteksi Granger, khususnya pengaruh suhu pada konsumsi gas alam.

Shuffle Test, TE Value

Model EvaluationEvaluasi · Model

MAE / RMSE / MAPE

Three complementary evaluation metrics: MAE as the basis for model comparison, RMSE sensitive to outliers, MAPE as an intuitive percentage error for cross-model comparison.Tiga metrik evaluasi komplementer: MAE sebagai basis komparasi, RMSE sensitif terhadap outlier, MAPE dalam persentase untuk perbandingan antar model yang intuitif dan mudah diinterpretasi.

3-Model ComparisonKomparasi 3 Model

05 — Results & AnalysisHasil & Analisis

Research FindingsTemuan Penelitian

Experimental results show Holt Method consistently outperforms for energy consumption, while BVAR and Seq2Seq LSTM complement for climate data. Causality analysis reveals significant bidirectional relationships between energy consumption and climate change. Hasil eksperimen menunjukkan Holt Method unggul konsisten untuk konsumsi energi, BVAR dan Seq2Seq LSTM melengkapi untuk data iklim, serta analisis kausalitas mengungkap hubungan dua arah yang signifikan antara konsumsi energi dan perubahan iklim.

Model Comparison — Energy ConsumptionKomparasi Model — Konsumsi Energi

VariableVariabelBest ModelModel TerbaikMAENotesKeterangan

Ind_Electricity ★

Holt

21,494

MAPE 4.50% — lowest MAE, best in class

Ind_Coal

Holt

34,196

Far better than BVAR (121,623) & LSTM (280,201)Jauh lebih baik dari BVAR (121.623) & LSTM (280.201)

HH_Natural_gasRT_Natural_gas

Holt

196

Absolute lowest MAE across all energy variablesMAE terendah mutlak di seluruh variabel energi

HH_Oil & HH_BiofuelsRT_Oil & RT_Biofuels

Holt

3,430 / 5,729

Consistently outperforms BVAR and Seq2Seq LSTMKonsisten unggul atas BVAR dan Seq2Seq LSTM

Com_ElectricityKom_Electricity

Holt

11,354

MAPE 4.78% — stable across commercial sectorstabil di seluruh sektor komersial

Model Comparison — Climate ChangeKomparasi Model — Perubahan Iklim

VariableVariabelBest ModelModel TerbaikResults & NotesHasil & Keterangan

Temperature ★

BVAR

MAE 0.11 — MAPE 10.95%; Seq2Seq LSTM MAPE very high (41.83%)

Emission_CO₂ ★

Seq2Seq LSTM

MAE 73,063,336 — MAPE 10.57%; better than BVAR (MAE 91,239,503)

Emission_GRK ★

BVAR

MAPE 10.64% — stable on consistently patterned GRK datastabil pada data GRK yang berpola konsisten

Granger Causality Test

Ind_Electricity → Emission_CO₂ lag 1 (F=12.39, p=0.0014) — industrial electricity consumption is the strongest causal predictor of carbon emissions.konsumsi listrik industri adalah prediktor kausal terkuat emisi karbon.

Ind_Electricity → Emission_CO₂ lag 2 (F=8.38, p=0.0016) — causal relationship persists up to 2-year lag.hubungan kausal bertahan hingga lag 2 tahun.

Emission_CO₂ → HH_Natural_gasRT_Natural_gas lag 5 (F=4.78, p=0.0066) — CO₂ emissions influence residential natural gas consumption in the long run.emisi CO₂ berpengaruh terhadap konsumsi gas alam rumah tangga jangka panjang.

Emission_GRK → Com_BiofuelsKom_Biofuels lag 5 (F=4.09, p=0.0128) — GRK emissions consistently influence commercial biofuel consumption.emisi GRK mempengaruhi konsumsi biofuel komersial secara konsisten.

HH_BiofuelsRT_Biofuels → Temperature lag 2 (F=5.13, p=0.0132) — demonstrates an energy-climate feedback loop mechanism.menunjukkan mekanisme umpan balik (feedback loop) energi-iklim.

Transfer Entropy

Temperature → Ind_Natural_gas lag 1 (TE=0.77, p=0.00) — temperature change's influence on industrial gas consumption is strongest.pengaruh perubahan suhu terhadap konsumsi gas industri paling kuat.

Temperature → HH_Natural_gasRT_Natural_gas lag 1 (TE=0.76, p=0.01) — temperature non-linearly influences residential natural gas consumption.suhu mempengaruhi konsumsi gas alam rumah tangga secara non-linear.

Ind_Coal → Emission_CO₂ lag 1 (TE=0.51, p=0.08) — information flow from coal consumption to carbon emissions detected.aliran informasi dari konsumsi batubara ke emisi karbon terdeteksi.

Ind_Oil → Emission_GRK lag 2 (TE=0.76, p=0.08) — industrial oil consumption flows information to GHG emissions.konsumsi minyak industri mengalirkan informasi ke emisi GRK.

Non-linear relationships undetected by Granger are confirmed by Transfer Entropy, revealing complex energy-climate dynamics.Hubungan non-linier yang tidak terdeteksi Granger terkonfirmasi Transfer Entropy, menunjukkan dinamika energi-iklim yang kompleks.

Main ConclusionKesimpulan Utama

This research successfully achieved all stated objectives. Holt Method proved consistently best for all energy consumption variables, with the lowest MAPE of 4.50%. BVAR excels for Temperature and Emission_GRK, while Seq2Seq LSTM is optimal for Emission_CO₂ (MAPE 10.57%). Causality analysis confirms bidirectional relationships that are not only linear but also involve complex non-linear dynamics between energy consumption and climate change in Indonesia. Penelitian ini berhasil mencapai seluruh tujuan yang dirumuskan. Model Holt Method terbukti konsisten sebagai model terbaik untuk peramalan seluruh variabel konsumsi energi, dengan MAPE terendah sebesar 4,50%. BVAR unggul untuk Temperature dan Emission_GRK, sementara Seq2Seq LSTM optimal untuk Emission_CO₂ (MAPE 10,57%). Analisis kausalitas mengkonfirmasi adanya hubungan dua arah yang tidak hanya bersifat linier tetapi juga melibatkan dinamika non-linear yang kompleks antara konsumsi energi dan perubahan iklim di Indonesia.

Analytic Dashboard of Energy Consumption and Climate Change in Indonesia Dashboard Analitik Konsumsi Energi dan Perubahan Iklim di Indonesia

About the SIECCI DashboardTentang Dashboard SIECCI

Energy ConsumptionKonsumsi Energi

Climate ChangePerubahan Iklim

Integrated ForecastingPeramalan Terpadu

Causality AnalysisAnalisis Kausalitas

SIECCI Analytics Dashboard

Research Insights & Policy Recommendations Insight Penelitian & Rekomendasi Kebijakan

Industrial Energy Transition Transisi Energi Sektor Industri

Commercial Sector Efficiency Efisiensi Energi Sektor Komersial

Household Energy Behaviour Kesadaran Energi Rumah Tangga

Data-Driven Monitoring System Monitoring Energi Berbasis Data

About the ProjectTentang Proyek

BackgroundLatar Belakang

Research Questions & ObjectivesRumusan Masalah & Tujuan

Research Dataset StructureStruktur Data Penelitian

Research MethodologyMetode Penelitian

Bayesian Vector AutoRegression

Sequence-to-Sequence LSTM

Double Exponential Smoothing

Granger Causality Test

Transfer Entropy

MAE / RMSE / MAPE

Research FindingsTemuan Penelitian

Granger Causality Test

Transfer Entropy