[eng] Length of stay has become a key variable for tourism management. However, estimation of tourists' length of stay implies the need to implement surveys of tourists. In addition to the significant costs associated to tourism surveys, their implementation over time is frequently interrupted, either due to budget issues or to methodological changes, and, consequently, monitoring average stay evolution is often difficult. In this paper, a new method for estimating tourists' length of stay from airport data is proposed and tested for the case of the Balearic Islands (Spain). Using regression analysis techniques, it is found how it is possible to replicate estimations for tourists' length of stay in the Balearic Islands obtained from surveys, for the different islands, different years, and also different nationalities; it is even possible to fill in the data where official statistics failed and obtain a homogenous estimation that targets multiple comparisons including the temporal one.