Data Science Desktop Survival Guide
by Graham Williams
Rattle Weather Dataset
Rattle provides a number of sample datasets. We can easily load them into Rattle. By default, Rattle will load the weather dataset.
We load the dataset in two simple steps
We can use this dataset for predictive modelling to predict if it might rain tomorrow (aka statistical classification and supervised learning), or to predict how much rain we might get tomorrow (aka regression analysis).
The dataset itself is quite small, consisting of 366 observations of some 22 variables. Rattle also provides a much larger weather dataset covering all of Australia. The weatherAUS dataset has 176,747 from 49 weather stattions, representing daily observations over many years.
We use the smaller dataset here for illustration, noting that typically, as data scientists, we often analyse significantly larger datasets.