Product quality during shipping by generating lane temperature and product temperature from models

A method for modeling a temperature in a thermal package in a lane of commerce, comprising:

• receiving weather data;
• selecting a weather model in accordance with an origin of a shipment, a destination of the shipment, a service level, and a date and a time of the shipment, the weather model being determined by a neural network machine learning algorithm that is trained using the weather data and lane data, wherein the neural network machine learning algorithm is determined using loss functions that calculate a difference between an estimated temperature and a validation set of data, and optimizers that use derivatives of multi-dimensional surfaces to determine variables weights, the loss functions and optimizers identifying an ideal structure of the neural network machine learning algorithm;
• determining an estimated lane temperature for the shipment using a selected weather model, wherein the estimated lane temperature is an estimate of a temperature it actually experienced by the thermal package between the origin and the destination;
• determining a product temperature using a package model in accordance with the estimated lane temperature;
• determining a packaging to be used for the shipment from the product temperature;
• and
• displaying a determined packaging in a user interface.