Gerardus Mercator (1512 – 1594) was a 16th century polymath who created the world’s first atlas, produced a novel way of plotting latitude and longitude that is still being used today (Mercator’s Projection), and wrote the first book on the italics font. In addition he produced fabulous maps of the polar regions, Europe, as well as a world map that were considered the standard for that time. Yet Mercator rarely ventured outside of a 50 mile radius in present-day Belgium that comprised where he spent the majority of his life. How this remarkable man born of humble origins accomplished this is wonderfully told in a book by Nicholas Crane called “Mercator: The Man Who Mapped the Planet“. But I want to describe how Mercator unwittingly was a data scientist whose methods should be emulated today.
First and foremost, Mercator got a broad-based education: mathematics, cosmology, geography, the fine arts, and theology. He picked up cartography after he had mastered those fields. Too many data scientists today concentrate on studying mathematics, statistics, computer science, or some combination of these subjects. Having the technical knowledge will only take you so far. It’s critical to have at least an understanding of either biology, chemistry, or finance to supplement your technical expertise.
Mercator the data scientist
Mercator acquired virtually all of his data second-hand. But he didn’t assume that the data were correct. He matched data from two or more sources, in essence carrying out his own quality control. When a new source of data contradicted what he believed, Mercator vetted it and did not hesitate to update his work, even if that meant discarding nostrums he had dearly believed. The lesson we can learn from this is that data quality is paramount. Brilliant analytic methods will not overcome data that are corrupt in some way.
One of Mercator’s greatest attributes was his ability to collaborate with other intellectuals of the day. This meant that someone might “scoop” him and publish a map before Mercator. But it was a risk he was willing to take in his quest for more knowledge. This was particularly germane to areas where others had far more domain area expertise. So as a data scientist you shouldn’t try to fly solo. Draw from a variety of colleagues or industry leaders.
Finally, Mercator was not reticent to switch course when circumstances dictated so. Mercator could have made a handsome living producing globes, which is what he started off doing. But Mercator saw a real gap in offering maps with quality data, standardized to a single format. As a result he threw himself into creating two-dimensional representations of the world. Data scientists can get stuck in a specific process that eventually becomes routine. So learn (& try!) new techniques. Look at your data through a different lens if you’re not getting acceptable models.
Be adaptable, never unflinchingly believe your results, collaborate, and acquire domain knowledge in the subject matter. And though you may not travel 50 miles beyond your computer, great analytic creations are still very much possible.