Master projects on language evolution and movement analysis.

Linguistic innovations are new forms of language that did not exist before. Many such innovations first appear online, for example on social media, from where they diffuse and permeate into everyday language. In her Master thesis, Patricia studied the spatio-temporal diffusion of linguistic  innovations on Spanish Twitter. From more than 50 million geo-tagged tweets she infers the network along which new linguistic innovations spread. Her results show that South American Twitter users tend to influence European users: innovations first appear in Argentina and Uruguay and are then picked up by mainland Spain.

The Master thesis is co-supervised by Prof. Carlota de Benito Moreno from the URPP Language and Space at the University of Zurich and Prof. Robert Weibel from the Department of Geography at the University of Zurich. 

Twice a year millions of migratory birds travel from their European breeding grounds to their sub-Saharan winter-habitat. On their way south, the birds have to overcome enormous obstacles, one of the most challenging being the passage of the Alps. For most birds, the Alpine peaks are too high to fly over. Instead, the birds take a detour, avoiding the Alps altogether, or navigate through lower valleys. In his Master thesis, Simon studies how birds adapt their migratory behavior when crossing the Alps. He explores the ecological barrier effect of the Alps and demonstrates the value of open passages for migratory bird conservation.

The Master thesis is co-supervised by Dr. Felix Liechti from the Swiss Ornithological Institute and Prof. Robert Weibel from the Department of Geography at the University of Zurich.

The Syntaktischer Atlas der deutschen Schweiz (SADS) surveyed more than 3000 individuals across German-speaking Switzerland on over 200 linguistic variants to capture the morphosyntactic variation in Swiss German in. In her Master thesis, Noemi Romano applied TESS, a Bayesian clustering method from evolutionary biology to the SADS data to infer population structure. Specifically, she tested three clustering models with different spatial assumptions: a non-spatial model, a spatial trend model with a spatial gradient, and a spatial full-trend model with both a spatial gradient and spatial-autocorrelation. Noemi’s results reveal five distinct morphosyntactic populations in Swiss-Germen, four of which correspond to traditional Swiss German dialect regions, and one of which corresponds to a Swiss base population.

The master thesis was co-supervised by Dr. Stéphane Joost from the Laboratory of Geographic Information Systems (LASIG) at the École Polytechnique Fédérale de Lausanne. 

For small birds, GPS tracking devices are too heavy and alternative tracking methods with lighter devices have to be used.  One of these, geo-location by light, estimates the migratory route from solar light-level intensities.  Unfortunately, geo-location is often coarse and unreliable, with positioning  errors  of  anything  up  to  hundreds  of  kilometers.   Recent  Bayesian  models  try  to constrain the route to plausible corridors:  they couple light-level measurements with information about the bird’s movement, such as the flight time and the ground speed.  However, these models routinely assume that the bird migrates steadily without influence from conditions en route. This assumption is hardly met.  Weather, and especially wind, are known to affect the ground speed and flight time, which in turn have an influence on probable migratory routes.  In his Master thesis, Mike adapted an existing Bayesian framework for inferring bird migration from light-level geo-locators.  Specifically, the model informs route inference with wind and activity measurements en route.  Mike applied the model to infer the migratory route of European Hoopoes (Upupa Epops).  

The Master thesis was co-supervised by Dr. Felix Liechti from the Swiss Ornithological Institute and Prof. Robert Weibel from the Department of Geography at the University of Zurich.

In his Master thesis, Luca developed an unsupervised, data-driven approach to identify different user types based on high-resolution human movement data. The data were collected in Australia  by Sygic, a smartphone navigation app. Luca’s approach identifies two distinct groups of users: ‘travelers’ who visit different areas with distinct, salient characteristics, and ‘locals’, who cover shorter distances and revisit many of their locations. Luca verified the  approach by relating user types to space use:  travelers and locals prefer to visit distinct, different locations in the Australian cities Sydney and Melbourne. 

The Master thesis was co-supervised by Dr. Martin Tomko from the Department of Infrastructure Engineering at the University of Melbourne and Prof. Robert Weibel from the Department of Geography at the University of Zurich.