**5. Conclusions**

The aim of this research is to show how to study the variations in energy demand and indoor temperature when using two different *actual* weather data sources with the purpose of analyzing if the data from the sensors of an *on-site* weather station (with high economic cost and maintenance) could be replaced by the data provided by a *third-party*. In this regard, it can be concluded that this research is not enough to make a general evaluation of the impact of using third-party *actual* weather data, but it has shown relevant variations in the energy demand and indoor temperature in the four test sites when both weather datasets are used in the simulations, especially for hourly criteria (used in calibration processes and in other applications such as model predictive control). The study also showed that for these types of building energy models, which employ dynamic infiltrations, wind speed's influence on the energy demand is relevant. The significant variations in the results lead us to make the recommendation for researchers to analyze in detail the impact on the building

energy models using *third-party actual* weather data before employing it. For example, this analysis methodology could be performed before making an investment into a weather station by installing a rented provisional one for a period of time. This way, the most influential weather parameters for a specific building energy model and a *third-party* weather provider could be determined, and with this information, an informed choice about which sensors are worth being purchased and installed can be made.

**Author Contributions:** E.L.S. supervised the methodology used in the article, performed the simulations and the analysis, and wrote the manuscript. G.R.R. and C.F.B. developed the methodology and participated in the data analysis. V.G.G. and G.R.R. developed the EnergyPlus models. A.P. provided resources for the study. All the authors revised and verified the manuscript before sending it to the journal. All authors read and agreed to the published version of the manuscript.

**Funding:** This project received funding from the European Union's Horizon 2020 research and innovation program under Grant Agreement No. 731211, project SABINA.

**Acknowledgments:** We would like to thank the National Technical University of Athens for providing the data of the H2SusBuild and administration building test sites located in Lavrion (Greece) and also Insero in the case of the Gedved school test site in Denmark.

**Conflicts of Interest:** The authors declare no conflicts of interest.
