Reproducibility and Replicability
One of the foundationl pillars of the scientific method is reproducibility. If an experiment or study can’t be reproduced, its results may illuminate a fluke or error as opposed to real scientific insight. There is a crisis of reproducibility in the geographic sciences for two main reasons. First, geographic experiments are almost impossible to create, and therefore most analysis relies on natural experiments. Second, the software and data used in one study may be obsolete, impossible to obtain, or rendered unusable by the time a replication study is carried out.
Learning how to code more complicated queries using SQL has made using QGIS that much more interesting. Being able to read queries and understand what the output should be is very exciting, and much more challenging than the visual layout of QGIS tools. Having more control over commands, and having my computer respond much faster than it would if I was rendering layer changes in real time, is going to be a handy tool for my thesis. For my senior thesis I will be working with a large dataset of river quality data. Being able to process the data, and handily keep a record of all changes and queries in an Atom document, will enable me to focus on data analysis as opposed to fumbling with Q or excel. SQL represents another stripped away layer of abstraction that I spoke about in my first blog post.