As we look at our Simulation, and analyze the problem from the unfortunate position of a failed simulation, there are elements that come clear, both in our structure, and the root cause of the the problem.
In the same way that a chef is dependent on the ingredients at hand, it is critical to the success of our simulation to have data that is accessible to our scripts. There are a number of resources available, but none are readily accessible; the inconsistent results in our first simulation are, in large part, due to the inconsistent parsing of the data.
This was a hidden effect; it wasn't until we spent time examining the data that we had assumed scraped cleanly into our script that the severity of the problem became clear. The data was from a reliable source, and detailed enough for analysis and simulation; however, it came in the form of a .pdf that could not be converted and scraped in a way that diligent work with regex could relieve. Unfortunately, there were over 3000 points of data on Isotopes, and that is a daunting task to hand-code into our script.
Thus, our fist assumption that we were simulating with accurate data was incorrect and difficult to correct.
Below is an example of the structure of the data, once it has been reliable input explicitly; the two-dimensional array is particularly useful for comparisons and the simulation logic.
No comments:
Post a Comment