India is the third-largest contributor to global energy-use and anthropogenic carbon emissions. India’s urban energy transitions are critical to meet its climate goals due to the country’s rapid urbanization. However, no baseline urban energy-use dataset covers all Indian urban districts in ways that align with national totals and integrate social-economic-infrastructural attributes to inform such transitions. This paper develops a novel bottom-up plus top-down approach, comprehensively integrating multiple field surveys and utilizing machine learning, to model All Urban areas’ Energy-use (AllUrE) across all 640 districts in India, merged with social-economic-infrastructural data. Energy use estimates in this AllUrE-India dataset are evaluated by comparing with reported energy-use at three scales: nation-wide, state-wide, and city-level. Spatially granular AllUrE data aggregated nationally show good agreement with national totals (<2% difference). The goodness-of-fit ranged from 0.78–0.95 for comparison with state-level totals, and 0.90–0.99 with city-level data for different sectors. The relatively strong alignment at all three spatial scales demonstrates the value of AllUrE-India data for modelling urban energy transitions consistent with national energy and climate goals.

印度是全球能源使用和人为碳排放的第三大贡献者。由于印度的快速城市化进程，该国的城市能源转型对于实现其气候目标至关重要。然而，没有基准城市能源使用数据集能够以与全国总量一致的方式覆盖所有印度城市地区，并整合社会经济基础设施属性来为此类转型提供信息。本文开发了一种新颖的自下而上加自上而下的方法，全面整合多个实地调查并利用机器学习，对印度所有 640 个地区的所有城市地区的能源使用 (AllUrE) 进行建模，并与社会经济基础设施相结合数据。 AllUrE-India 数据集中的能源使用估算是通过与三个层面报告的能源使用情况进行比较来评估的：全国范围、州范围和城市级别。全国汇总的空间粒度 AllUrE 数据与全国总数显示出良好的一致性（差异 <2%）。与国家级总量比较的拟合优度范围为0.78-0.95，与不同行业的市级数据比较的拟合优度范围为0.90-0.99。所有三个空间尺度上相对较强的一致性证明了 AllUrE-India 数据对于模拟符合国家能源和气候目标的城市能源转型的价值。