Nitrogen response (N-response) curves are tools used to support farm management decisions. Conventionally, the N-response curve is modeled as an exponential function that aims to identify an important threshold for a given field: the economic optimum point. This is useful to determine the nitrogen rate beyond which there is no actual profit for the farmers. In this work, we show that N-response curves are not only field-specific but also site-specific and, as such, economic optimum points should be calculated locally. In addition, we propose a methodology that allows deriving N-response curves automatically instead of using parametric functions. Thus, we obtain a specific non-parametric N-response curve for each 10x10m cell of the field. First, we train a convolutional neural network called Hyper3DNetReg using data collected during the early stage of the winter wheat growing season (March) to predict yield values during the harvest season (August). This network models the behavior of the field under different environmental and terrain conditions. Then, we use the trained prediction model to obtain an N-response curve per cell by simulating what would be the yield response given a range of nitrogen rate values between 0 and 150 pounds per acre (lbs/ac). Results show that the shape of the N-response curve depends greatly on the region of the field from which it was calculated. What is more, our analysis allowed us to find regions of the field that are relatively insensitive to the nitrogen rate applied, which was also observed empirically. Finally, we show that site-specific economic optimum points can be easily calculated using the estimated first derivative of their corresponding generated N-response curves. Related work will address the problem of generating prescription maps that merge the calculated site-specific economic optimum points while also while minimizing the overall fertilizer applied and the number of jumps between consecutive cells’ nitrogen rates.