I require an intersting dependent variable sourced from a respectable website (ie - UK Office of National Statistics, etc) as well as 2-4 independent variables sourced from sites which attempt to explain the dependent variable. They don't have to actually explain the dependent variable but at least sound like a valid hypothesis.
Each variable requires a minimum of 40 years time-series data. The data should have a high degree of variability (ie difficult to model or predict) - but above all be interesting.
And naturally, the independent variables should make sense - ie be the most realistic explainers of the dependent variable.
NOTE: the data selected CANNOT relate to consumer expenditure.
An article that uses time series data of forty years and makes some interesting inferences is dissected.