Meteorological data

This dataset provides comprehensive meteorological time-series data for two major Indian cities, Jaipur and Kanpur, structured to facilitate robust climate analysis, environmental modeling, and machine learning research. The collection comprises four CSV files: two for each city, differentiated by the presence or absence of missing values. For Jaipur, both "with missing value" and "without missing value" datasets contain 575 records across six primary atmospheric variables: Dew point, Cloud cover, Precipitation, Temperature, Wind direction, and Wind speed. The Kanpur datasets follow a similar design but include an additional 'date\_time' column, representing concatenated timestamps, and extend to 575 records with seven columns in total. The inclusion of both raw (with missing values) and preprocessed (imputed, without missing values) variants enables users to benchmark data cleaning, imputation, and gap-filling algorithms, offering valuable real-world scenarios where sensor data is incomplete. All variables are provided in a consistent numerical format, with the imputed files utilizing float datatypes for improved compatibility with statistical and computational methods. The temporal granularity and spatial specificity of the data make it well-suited for urban climate studies, high-resolution forecasting, and analysis of atmospheric dynamics in Northern India. While the units for each variable (e.g., temperature, wind speed) follow standard meteorological conventions, users are advised to consult supplementary documentation or the data authors for precise measurement details, as direct unit annotation is not present within the files. The 'date\_time' field in the Kanpur datasets warrants special attention due to its composite format, and may require parsing for fine-grained temporal analysis. The data capture typical meteorological variation as observed by regional weather stations, including natural fluctuations, diurnal cycles, and seasonal transitions, while also realistically reflecting common challenges in environmental monitoring, such as sporadic missing values arising from sensor outages or data transmission errors. By providing both incomplete and fully-imputed datasets, this resource supports the development and evaluation of robust analytical pipelines in climate informatics and related disciplines. Researchers can leverage these datasets for tasks ranging from baseline climate characterization and anomaly detection to the training and validation of data-driven predictive models. Ultimately, the structured format, practical relevance, and dual-variant design position this dataset as a valuable open-access asset for atmospheric scientists, data engineers, and interdisciplinary researchers interested in advancing methods for meteorological data analysis, urban climatology, and environmental data science.