Making sense of the data...Data analysis

[osu]

Dear Experts,
I have data for a 12 months study done on 2 wastewater (Plant A and Plant B—Final Effluent and Discharge Point) treatment plants. Physiochemical (13 parameters tested) and Microbiological(bacterial) data were collected. Also there is the recommended standards against which all parameters recorded are checked to see if the treatment plants complied. What statistic analysis will suit the study. Is there also the possibility of looking at the trends of the parameters measured?, examine the relationships between concentrations of bacterial indicators and a variety of physiochemical indicators? comparison of the two plants based on their physiochemical and microbiological data?, statistical significant?, determine the level of interaction between the predictor variables, which parameter indicators had the strongest influence, and the total amount of variation that could be explained by the study. This are just some of the few ideas I have but I love to hear your expert views. So far just the descriptive statistics have been applied.

Some of the other things I hope to achieve are
to evaluate the compliance of the treatment plants to the set standards
to evaluate the efficiency of the treatment plants
to compare two outflow points (final effluent and discharge point)
to compare the two wastewater treatment plants to determine which plant is more functional

Thank you all

 

[A likert]

OSU,
Do you have monthly data, or is it more frequent? Comparing to the standard is easy. Comparing the two outflow points is also easy. If the data were collected on the same day you might expect a correlation between the measurements. A scatter plot will give you a good idea if there may be a correlation. Linear regression will give you numbers. I'm not sure what you mean by efficient or functional.
You can look for trends between parameters, but if you only have 12 independent measurements (assuming the two outflow points are correlated), then you need to be careful about interpreting anything that looks statistically significant. If you make 20 comparisons, you should get one significant at the 95% CL by chance.