Emilio Casetti, who has died at the age of 89, three weeks short of his ninetieth birthday, made important contributions to both geography and regional science. He was honored at Ohio State with the designation Distinguished Research Scholar in 1992, and in 1994 the Association of American Geographers bestowed upon him the Honors Award, the organization’s most prestigious recognition for research excellence. He was also an unforgettable person.
He acquired his bachelor’s degree in law at the Sapienza University in Rome at the early age of 20 and earned a doctorate from the University of Rome in three years. He then practiced law in Rome and had begun work towards a doctorate in geography at the same university before leaving to do a Master’s at McGill. Afterwards, he and his wife Gabriella spent several years in a remote part of Saskatchewan teaching in a very small rural school. He did not have to be pushed to regale the listener with hair raising tales of the winters there, all told with his standard deadpan grin. Quite what had led him in the direction of geography is unclear, but by 1964 he had graduated from Northwestern deeply steeped in the quantitative methods for which the Geography Department there was, at that time, notable. His first appointment was at Toronto but in 1966 he moved to the Department at Ohio State, where he would remain until his retirement.
He would become an important influence in graduate work in the Department and a renowned advisor. What made him so effective was his ability to see the unique strengths of each of his students, help them recognize it, and then allow them to deploy that strength to its fullest advantage. Not surprisingly, his students – and nine of his twenty-three doctoral students were, significantly for the time, women – have gone on to successful careers of their own, each in a direction that no doubt Emilio saw early on. They speak affectionately of him, and of the time he decided that they must call him ‘Emilio.’ He cared deeply for them and would defend them vigorously against pettiness.
His work was notable for its combination of simplicity, power and imagination. He himself thought that his development of what he called the expansion method was his major contribution. What this involved was taking some relationship, like Fourastié’s model of sectoral shift in a national economy, describing it with a regression model and then expanding it by setting the coefficients as functions of other pertinent variables, like, perhaps, the date at which an economy took off. This would then create a set of relationships that could stimulate further investigation. This would be a precursor to spatial regression models where the coefficients are a function of absolute or relative location.
Emilio’s expansion method led to a number of dissertations at Ohio State in the 1980s and through the 1990s. As he came to view things, however, it was not simply a method but a new paradigm for research, one that challenged scientific geography – of which he was a strong proponent – to entirely rethink the nomothetic (or law-seeking) enterprise. Instead of assuming that parameter instability was an aberration due to model misspecification or systematic biases in data and error terms, Emilio came to see the search for contextual variation in causal processes as part-and-parcel of the explanatory effort itself. For any model, the question of parameter variation opened up new questions, ones often more important than the first order explanatory questions tested in what he called the ‘initial model’.
Though he was well known for the expansion method, he was also the early inventor of what later became popularized as geographically-weighted regression (GWR). In his ‘drift-analysis-of-regression parameters’, or DARP, he constructed a grid across a set of spatially distributed observations and proceeded to run separate cell-by-cell regressions, weighting all observations according to an exponential distance-decay function from the cell’s central point. The result was a moving lattice of regression analyses that varied by the degree of information content from nearby data points, thus producing a series of ‘local’ regressions for which investigators could track and map parameter variation. Depending on the exponents determining the distance-decay of information from near or far away cases, the models could adjusted to be ‘regional’ as opposed to local. The insight led him to argue that spatially weighted regression is not a special case of regression; rather, analyses that don’t explicitly involve weighting are equivalent to assigning each case a value of 1, which is the norm for most models. Hence DARP, or geographically-weighted regression, as it became popularized, was the more general case. It is a sidebar in the history of geography that DARP preceded a now popular GWR by a decade and furthered the development of quantitative geography away from its search for universal laws to something much more sensitive to issues of context.
As a person he was a curious mix of strong views strongly held, and a sweet innocence. He was dedicated to his research but found lots of time for other things, the most notable of which were cats. Emilio was a cat lover, par excellence; not one who hoards them but who, along with his wife, took in strays in sufficiently modest numbers where they could develop an attachment to them. And he was attached. An invitation to dinner would be accepted but on his insistence, he and his wife would arrive in separate cars so that if one was involved in an accident, the cats would not go wanting. He was a very kind man and not just to cats. There are many stories of Emilio’s consideration of those students who had trouble with the quantitative methods that he taught.
There are other interesting stories about him. As a young adult growing up in postwar Italy, Emilio developed a strong aversion to traditional social institutions that he saw as oppressive and antithetical to progress. He joined the Italian communist party, but after several years concluded that the party itself reproduced the social hierarchies that he opposed. Many years later when he applied for a U.S. green card, he had to respond to a question about whether he had ever been a communist party member. He was honest in his response, and attached a statement describing how and why he was no longer a communist: nepotism and cronyism. He must be the one and only person who was given a green card after openly admitting that he had been a communist party member!
In the late 60s, he acquired, quite cheaply, one of the earlier analog computers and put it in his basement. This meant that he did not have to submit his jobs to the computing facility at the university, as was the custom at that time. The problem was, that it was huge. Being shown it was like walking down the book shelves in a library. Emilio was also an early adopter of desktops and laptops, and once they became available, getting rid of his heap of metal – whose only real purpose was in solving partial differential equations – was a nightmare. Through it all one could see Emilio’s usual sense of humor and self-deprecation.
He liked people who he judged to be without guile and who he perceived to be honest. He will be deeply missed by those touched by his unique blend of kindness, considerateness and innocent disbelief that you might disagree with him. His life will endure through the major contributions that he made to spatial-quantitative geography and through the affectionate recollections of his students.