Financial institutions require bankable datasets to guarantee investments in large-scale solar energy projects, reducing the overall uncertainty of energy yield estimates. Long-term satellite-based data is mostly available globally; however, their systematic errors and bias should be removed by integrating quality-checked ground measurements if available. In this study, a quality assessment was done to detect erroneous/ missing data points using several quality-control tests. In METU NCC, global horizontal irradiation (GHI), direct normal irradiation (DNI), and global tilted irradiation (GTI) were recorded since 2010, 2013 and 2016, respectively. Physical threshold and quality envelope tests revealed that lower-than- expected GHI values were measured in some periods. Thus, erroneous GHI data was estimated using the Erbs model and measured DNI data. While up to 2 hours of missing GHI data were filled by linear interpolation, more extended missing data were filled by estimated GHI. Both measured and constructed GHI data were compared with satellite-based GHI data downloaded from the Photovoltaic Geographical Information System (PVGIS) for the time period 2010-2016. The results indicated that the average relative root-mean-square error (rRMSE) of daily total GHI reduced from 34.63% to 17.77% after the data filing process. The rRMSE decreased to 8.78% for the annual mean daily total GHI data. Additionally, GTI was estimated using the isotropic sky-diffuse model, and it was compared with measured GTI. RMSE of daily total estimated GTI was 22.41%, whereas satellite-based GTI had an RMSE of 20.81%. Finally, energy yield was estimated for 1 MWp solar photovoltaic (PV) plant using satellite-based and estimated GTI.