Our model is broken into three different stages: pre-processing, computation, and visualization. Pre-processing removes outliers, impossible values, and otherwise cleans the data. For example removing altitudes from flight data above the aircraft or legal service ceiling, or removing data from impossible locations.
Computation includes combining the flight and weather data and running it through the Schmidt-Appleman formulas to determine the likelihood of a contrail forming. This phase of the model proved to be the hardest and the challenges will be discussed further later. The final part of the model was visualization, for this we used standard matplotlib libraries.
Implementing the Schmidt-Appleman criterion turned out to be the most difficult part of the implementation. This means we were unfortunately unable to implement this in the given timeframe. The Schmidt-Appleman criterion requires knowledge of at least Water vapor partial pressure and temperature at the site of potential formation. This is data we don’t have access to nor do we have time or processing power enough to simulate it. As such we can only leave you with recommendations for continued research in this field.
