Based on my own observations, there is a great correlation between the daily weather and the daily income of the companies which offer outdoor entertainment services. This amount depends on the average temperature, the rainfall etc. To get know the relation between the weather conditions and the income as precise as possible can be very helpful to forecast the daily incomes. In this paper I describe and examine some widely used related statistic and data mining methods to discover their accuracy. I focused especially on the nearest neighbor and the artificial neural network. As a result I didn’t get a clearly answer which of the two is more effective but by changing their setting details I managed to reduce the error pretty much for both of them: the optimal artificial neural network consisted of one hidden layer with five neurons in it. It was trained with Levenber-Marquardt backpropagation algorithm and the transfer characteristic was tangent sigmoid for the hidden layer and for the output layer as well.