Modelling the occurrence of the Usutu virus using opportunistic dataDescription:
Natuurpunt is well-known in Flanders for organizing its bird-counting weekends. Every citizen is allowed to count the birds he observes during the weekend and to pass this information to Natuurpunt (http://waarnemingen.be). Citizen science is an umbrella term for such events and others where citizens are involved in scientific research. The data that is voluntarily collected by individuals with different levels of expertise and in an unstandardized manner is also called 'opportunistic data'.
Opportunistic data suffers from several biases making the statistical analysis very challenging. Examples of error and bias include wrong identifications of species, detection bias, and sampling bias among others [1]. In this thesis, we will focus on the problem of sampling bias. This occurs when some samples are more likely to occur than others, resulting in an uneven distribution of samples over time and space (e.g. observations will often be clustered around urban centres or roads because they are easily accessible to volunteers).
In this thesis the possibility can be studied to use opportunistic data to model the occurrence of the Usutu virus that caused several deaths among birds in Belgium and of which the first cases in Belgium were observed in 2016.
The student will closely collaborate with Natuurpunt. Firstly, one or more data sets are extracted from http://waarnemingen.be The student gives a literature review on techniques that can be used to model the data. The general aim is to apply a model on the dataset from Natuurpunt and obtain predictions of occurrence probabilities.
References:
[1] Bird, T. J. et al. Statistical solutions for error and bias in global citizen science datasets. Biol. Conserv. 173, 144-154 (2014).
Promotor(s): Stijn Luca and Marc Herremans (Natuurpunt) More info: stijn.luca@ugent.be Background: All
|