Abstract For managers of coastal areas, data and statistics on the usage and appreciation of nature are important. Utilization of social media platforms, such as the photo-sharing website Flickr, is a potential trend. We propose a unique strategy based on machine learning (image analysis) and convolutional neural network (CNN) for assessing cultural ecosystem services (CES) by collecting, and interpreting 29,000 photographs from the Lithuanian costal area. The most often represented CES categories were landscape appreciation and social recreation, which reflects the evident benefits. Similarly, historical monuments and environmental enjoyment were well represented. Engagements with CES followed distinct geographical and temporal patterns that were related to user behavior and reflected the infrastructure features of various places along the Lithuanian coast. Due to the data's extensive spatial coverage and high spatio-temporal resolution, this technique is suitable for finding CES hotspots/cold spots and, despite limits, holds promising potential for monitoring the impact of management actions on CES provision. Our study demonstrates how analyzing large amounts of digital photographs expands the analytical toolbox available to researchers and allows the quantification and mapping of CES at large geographical scales.