To do the experiment in Azure Machine Learning Studio we used a softward called orange. Orange gave us the same potentiality as Azure Ml, but it helped us run our experiments faster locally in our computer. We didn't have to communicate with the web. As soon as we had the desired results we replicated the whole process in Azure Machine Learning Studio. The advantage of the Azure platform is the deployment of algorithm, which was of course key for our assignment. Let's describe briefly what we did in Azure ML. First of all we loaded the data. We started by doing some preprocessing. Specifically we did forward selection to pick the variables that best predict the price of the house. We decided to use 7 variables. We trained our algorithm using 80% of the data as train data and test on the remain 20% of the data. In the orange we compared the accuracy of different algorithms. The best one to use was random forest. So we used that one for our experiment in Microsoft Azure. We were able to achieve a 81,24 Coefficient of determination. We then update our predictive model and we are ready to deploy the model.