# CG Q0 # Read the data file bikeshare.csv into R and name the object bikes.

####### As usual, don't forget the strings = T argument.

# CG Q1 # Run str() on the bikes data frame to see that

###### there are quantitative variables and to confirm that the

###### strings = T argument did in fact allow character data

###### to be read in a a factor.

# Question 2 - Simple linear model.

# CG Q2a # Build a linear regression model that has

####### count as the response and

####### the weather situation variable as predictor.

####### Name your fitted model simplefit.

# CG Q2b # Use the summary() function on simplefit

###### to access the results of the regression.

# CG Q2c # Print the coefficient for the wet weather situation.

####### Use coef(simplefit) followed by the name of that

####### coefficient in quotes inside square brackets.

####### For example, coef(simplefit)["weathersitcloudy"] prints

####### the coefficient for the cloudy weather situation.

# CG Q2d # Using the regression output, determine how the ride count

####### for wet days compares to the ride count on clear days.

####### If it's higher on wet days, type paste("higher")

####### If it's lower on wet days, type paste("lower")

# CG Q2e # Find the R-squared for the regression.

####### In your calculation, use simplefit$deviance and

####### simplefit$null.deviance and not the numbers printed

####### in the summary output for these.

# Question 3 - Linear model with multiple predictors.

# CG Q3a # Run a linear regression using ride counts as the response

###### modeled by the weather variables weathersit, temp, hum, and windspeed.

###### Use an additive model (don't model interactions or anything fancy).

###### Name your fitted model ridefit.

# CG Q3b # Use summary() on your fitted model to print the results.

# CG Q3c # How does expected ride count change with an increase in temperature?

####### Print the coefficient for temp in the ridefit model

####### using the same strategy used in Q2c.

# CG Q3d # Based on the temp coefficient, do we expect ride count to increase

####### or decrease by 156 rides with an increase in temperature?

####### If ride count is expected to increase with temperature,

####### type paste("increase"). Otherwise paste("decrease").

# CG Q3e # Find the R-squared for the ridefit regression.

####### In your calculation, use ridefit$deviance and

####### ridefit$null.deviance and not the numbers printed

####### in the summary output for these.

# Question 4 - log-linear model.

# CG Q4a # Now, run a linear regression using log ride counts as the response

###### modeled by the weather variables weathersit, temp, hum, and windspeed.

###### Use these same predictors as ridefit and name this new model logridefit

# CG Q4b # Use summary() on your fitted model to print the results.

# CG Q4c # How does expected log ride count change with an increase in temperature?

####### Print the coefficient for temp in the logridefit model

####### using the same strategy used in Q2c.

# CG Q4d # Based on the temp coefficient from log ridefit,

####### do we expect ride count to increase

####### or decrease with an increase in temperature?

####### If log ride count is expected to increae with temperature,

####### type paste("increase"), otherwise paste("decrease").

# CG Q4e # Now print the multiplicative effect of temperature

####### on expected ride count by wrapping the line of code

####### from Q4c in the exp() function.

# CG Q4f # Find the R-squared for the logridefit regression.

####### In your calculation, use logridefit$deviance and

####### logridefit$null.deviance and not the numbers printed

####### in the summary output for these.

# Question 5 - log-log model

# CG Q5a # Now, run a linear regression using log ride counts as the response

###### modeled by the weather variables weathersit, log(temp), hum, and windspeed.

###### Note the log() on the temp predictor.

###### Name your model loglogfit.

# CG Q5b # Use summary() on your fitted model to print the results.

# CG Q5c # How does expected percent ride count change with a

####### 1% increase in temperature?

####### Print the coefficient for log(temp) in the loglogfit model

####### using the same strategy used in Q2c.

# CG Q5d # Based on the log(temp) coefficient from loglogfit,

####### do we expect a percent increase or decrease in

####### ride count for each 1% increase in temperature?

####### If a percent increase is expected, type paste("increase")

####### otherwise paste("decrease")

# CG Q5e # Find the R-squared for the loglogfit regression.

####### In your calculation, use loglogfit$deviance and

####### loglogfit$null.deviance and not the numbers printed

####### in the summary output for these.

# Question 6 - Let's practice making predictions of the response variable.

# CG Q6a # We'll make predictions using all 3 competing models for

####### a wet day that is 25 degrees, has 50% humidity and windspeed of 5.

####### Make a data frame of these values and call it newdata.

# CG Q6b # Now use ridefit model to predict ride count using

####### the predict() function with your newdata object.

# CG Q6c # Now use the logridefit model to predict log ride count using

####### the predict() function with your newdata object.

# CG Q6d # To put the prediction of the log ride count back onto the raw

####### ride count scale, wrap the line of code from Q6c in the exp()

####### function to "undo" the log transformation.

# CG Q6e # use the loglogfit model to predict log ride count using

####### the predict() function with your newdata object.

# CG Q6f # To put the prediction of the log ride count back onto the raw

####### ride count scale, wrap the line of code from Q6e in the exp()

####### function to "undo" the log transformation.

answer all of these