26 lines
1.4 KiB
Plaintext
26 lines
1.4 KiB
Plaintext
---
|
|
title: "Project Work"
|
|
---
|
|
|
|
A considerable of time will be dedicated to project work allowing you to apply the skills you have learnt during the summer school.
|
|
|
|
We encourage you to bring your own problem (e.g. something related to your own research) and try to build a small software package for that problem.
|
|
|
|
Moreover, we will provide alternative problems to be solved:
|
|
|
|
## Project: Analyzing Spatial Data
|
|
|
|
In this project, we propose to write a package that allows for the statistical analysis of spatially indexed data.
|
|
|
|
Before starting, you might look at an data example. To this end, download [TemperatureData.jld2](TemperatureData.jld2) and read this data via
|
|
|
|
``` julia
|
|
|
|
using JLD2
|
|
@load "TemperatureData.jld2" ## potentially correct your path
|
|
```
|
|
|
|
This will give you data frames with information on weather stations in the Netherlands, summer temperature means (in $0.1^\circ\mathrm{C}$) and summer temperature maxima (in $0.1^\circ\mathrm{C}$). It is reasonable to assume that the temperature data are temporally independent. Furthermore, assuming a normal distribution at each station is quite common for this type of data.
|
|
|
|
A reasonable package could for instance contain functions to estimate means and standard deviations station-by-station, fit linear models for $\mu$ or $\log(\sigma)$ (with, e.g., geographic coordinates as covariates), or, provided that $\sigma$ is constant over space, a linear model for the data themselves (with repetitions.) \$
|