The dataset FloridaLakes includes information on lake water in Florida. We want to build a model to predict AvgMercury, which is the average mercury level of fish in the lake. Start with a model including the following four explanatory variables: Alkalinity, pH, Calcium, and Chlorophyll. Eliminate variables (and justify your decisions) and comment on how the model changes. Decide which model you believe is best using only these variables or a subset of them. Give the model you believe is best and explain why and how you chose it as the best model.

Dataset FloridaLakes

This dataset describes characteristics of water and fish samples from 53 Florida lakes. Some variables (e.g. Alkalinity, pH, and Calcium) reflect the chemistry of the water samples. Mercury levels were recorded for a sample of large mouth bass selected at each lake. Source: Lange, Royals, and Connor, Transactions of the American Fisheries Society (1993)
ID .................................An identifying number for each lake
Lake .............................Name of the lake
Alkalinity .....................Concentration of calcium carbonate (in mg/L) pH Acidity
Calcium .......................Amount of calcium in water
Chlorophyll .................Amount of chlorophyll in water
AvgMercury .................Average mercury level for a sample of fish (large mouth bass) from each lake
NumSamples ..............Number of fish sampled at each lake
MinMercury ................Minimum mercury level in a sampled fish
MaxMercury ...............Maximum mercury level in a sampled fish
ThreeYrStdMercury ....Adjusted mercury level to account for the age of the fish
AgeData ......................Mean age of fish in each sample