Electromobility is a key technology to decarbonize transportation and thereby avoid the worst impacts of anthropogenic climate change. To power such vehicles when away from their home or depot, public charging infrastructure is required which can be split into enroute and destination charging. We define the latter as charging events that occur while users are busy with other activities. To fulfill this purpose, chargers need to be placed in locations where people spend time. This paper introduces a novel approach to do so based on a neural network trained on several thousand public charging stations in Germany. Within the training sample, the approach is able to predict how much energy was charged per station and day with an R2 of 0.61 for the training set and a RMSE of 13 kWh/day. Using the network, we predict utilization across urban, suburban and industrial areas in Europe and make those predictions available through an easy-to-use web interface. It is further possible to perform predictions and, thereby, extrapolate the learnings from Germany to any country with sufficient OpenStreetMap data. The introduced holistic methodology with its prediction and visualization phase is a first-of-its-kind by applying large-scale measured charging data to the placement problem while being usable in areas which have not yet rolled out electromobility.