Data scientists build models of the real world using 1s and 0s. Model Railroaders build models of the real world using plastic and metal. In the end, they’re both models and Model Railroaders have been at it way longer than we DS have. Let’s look at parallel concepts like overfitting versus the 10 foot rule, synthetic data versus prototype freelancing, or assumptions versus modeler’s license and see what lessons from other realms of model building we can bring home to DS.
I'm Peter (he, him), a Geographer and Data Scientist based in New York City. I combine a deep domain expertise in geoinformatics and economic geography with technical skills in programming, machine learning, NLP, among others. I'm working to create 'Big Social Science'.