Predicting Innovative Cities Using Spatio-Temporal Activity Patterns.

Understanding cities’ complexity is essential for correctly developing public policies and urban management. Only some studies have attempted to relate the activity carried out by city inhabitants with the macro characteristics of a city, mainly its capacity to innovate. In this study, we seek to find those features that allow us to distinguish between an innovative city from those still on the way to becoming one. To carry out this analysis, we have the activity patterns decomposition obtained through geo-tagged social media digital traces and their respective innovation index for more than 100 cities worldwide. The results show that it is possible to predict the city’s innovative category from their activity patterns. Our model achieves an AUC = 0.71 and a KS = 0.42. This result is significant because it allows us to establish a relationship between the activities carried out by people in the city and their innovation index, a characteristic given for the capacity and development of cultural assets, infrastructure, and the quality of markets.