A Multi-Feature Dataset of Daily Electricity Demand with Meteorological, Holiday, and Special Event Attributes for Mirpur and Gulshan Zones of Dhaka, Bangladesh (2024–2025)

This dataset contains 597 daily electricity demand records covering the continuous period from 1 January 2024 to 19 August 2025 for two representative Electricity Distribution Service Areas (EDSAs) in Dhaka, Bangladesh: Mirpur (23.8223°N, 90.3654°E), a predominantly residential area, and Gulshan (23.7925°N, 90.4078°E), a predominantly commercial area. Electricity demand data were obtained directly from the Dhaka Electric Supply Company PLC (DESCO) with official authorization. The two zones, about 5.4 km apart, were selected to enable direct comparison of residential and commercial consumption under near-identical environmental conditions. The data are provided as two CSV files (MirpurDataset.csv and GulshanDataset.csv), each with 597 rows (one per day) and 13 identical columns, prepared using the same collection, preprocessing, feature engineering, and quality-checking pipeline. The target variable, Total Demand, is the total daily electricity demand in MWh, obtained by summing DESCO's 24 hourly demand values for each area. Daily demand ranged from 2,340 to 6,608 MWh in Mirpur and from 2,022 to 6,332 MWh in Gulshan. Fewer than 1% of demand observations were missing and were imputed as the average of the preceding and following days. Each record includes six daily weather variables — maximum, minimum, and mean temperature (°C), rainfall (mm), maximum wind speed (m/s), and sunshine duration (hours) — retrieved from the Open-Meteo historical archive for a reference point between the two areas and joined by date. Calendar features comprise the day of the week; one of seven local climatic seasons (Winter, Spring, Summer, Monsoon, Late Monsoon, Autumn, Late Autumn); holiday flags covering weekly holidays (Friday–Saturday), public holidays, and government-declared special holidays, annotated from official Government of Bangladesh calendars; and the names of major national, cultural, and religious events (e.g., Eid-ul-Fitr, Durga Puja, Pohela Boishakh), with overlapping events separated by "|" and "None" otherwise. The dataset supports electricity load forecasting, energy demand modeling with exogenous variables, holiday and festival effect analysis, smart-grid analytics, explainable AI research, and comparative residential–commercial energy studies.本数据集包含597条每日电力需求记录，覆盖孟加拉国达卡市两个典型配电服务区（Electricity Distribution Service Areas，EDSA）从2024年1月1日至2025年8月19日的连续观测时段：米尔布尔（Mirpur，23.8223°N，90.3654°E），以居住功能为主；古尔尚（Gulshan，23.7925°N，90.4078°E），以商业功能为主。电力需求数据经官方授权，直接取自达卡电力供应有限公司（Dhaka Electric Supply Company PLC，DESCO）。两个区域相距约5.4公里，可在近乎一致的环境条件下直接对比住宅与商业用电模式。 数据集以两个CSV文件形式提供，分别为MirpurDataset.csv与GulshanDataset.csv，每个文件包含597行数据（每日对应1行）与13个完全一致的字段，均采用统一的采集、预处理、特征工程与质量校验流程构建。目标变量为“总需求”（Total Demand），单位为兆瓦时（MWh），通过累加对应区域达卡电力供应有限公司的24小时逐时需求值计算得到。米尔布尔的日需求量区间为2340至6608 MWh，古尔尚为2022至6332 MWh。缺失的需求观测值占比不足1%，采用前后两日的平均值进行插补处理。 每条记录包含6项每日气象变量：最高、最低与平均气温（单位：摄氏度）、降雨量（单位：毫米）、最大风速（单位：米/秒）以及日照时长（单位：小时），数据取自两区域之间参考点的Open-Meteo历史气象存档，并通过日期字段完成关联。日历特征包括：星期几、7个本地气候季节（冬季、春季、夏季、季风季、晚季风季、秋季、晚秋季）之一、覆盖每周休息日（周五至周六）、公共假期与政府宣布的特别假期的假期标记（数据源自孟加拉国政府官方日历），以及主要国家、文化与宗教节日名称（如开斋节Eid-ul-Fitr、杜尔迦节Durga Puja、孟加拉新年Pohela Boishakh），多个重叠节日以“|”分隔，无节日时标注为“None”。 本数据集可支撑电力负荷预测、含外生变量的能源需求建模、假期与节日效应分析、智能电网分析、可解释AI（Explainable AI）研究以及住宅与商业用电对比研究等方向的工作。