In recent years, demand for databases of high quality weather observations has intensified. While there are many causes for this, increases in the number of climate studies and in the number of businesses employing weather risk management methods are two principal causes driving this increased demand. In response to this need, domestic and foreign governments, academic institutions and a few businesses have made efforts to improve the quality and access to such databases. This paper describes a comprehensive global weather observation database developed by Weather Source, LLC. The Weather Source database (WSDB) consists of daily, hourly and special weather observations from tens of thousands of weather stations from across the globe. The period of record for these observations in most cases covers the period from station inception to real-time if available. A series of quality control and data improvement methods are employed on the data to increase quality; these include both meteorological consistency and statistical methods which perform various functions. These functions include: identification of errors, production of high quality estimates to correct errors and fill missing values, and data homogenization (adjustments for observation discontinuities and long-term trends). Discussed herein is a description of the WSDB data, data improvement methods and data access options.