icecream: Icecream sales and temperature

In lgrdata: Example Datasets for a Learning Guide to R

Description

A synthetic dataset on weekly ice cream sales in two locations in Amsterdam, along with air temperature. The idea is that the ice cream salesman first sold icecream in 'Oosterpark', and decided to move shop to the 'Dappermarkt' the year after. Did sales improve? This dataset can be used to show that naive conclusions from simple linear model fits can be misleading, and that the use of covariates (here, air temperature) can change conclusions about effects.

Usage

icecream

Format

A data frame with 40 rows and 3 variables:

temperature

double Air temperature (C)

sales

double Icecream sales per week (in local currency)

location

factor Either 'Dappermarkt' or 'Oosterpark'

Examples

data(icecream)

# Linear model, temperature as covariate
fit_ice <- lm(sales ~ temperature*location, data=icecream)

# Try to guess from coefficients where the sales were higher:
summary(fit_ice)

# What about now?
with(icecream, plot(temperature, sales, pch=19, col=location))
legend("topleft", levels(icecream$location), fill=palette())

Example output

Call:
lm(formula = sales ~ temperature * location, data = icecream)

Residuals:
    Min      1Q  Median      3Q     Max 
-63.018 -32.222   3.449  28.157  83.390 

Coefficients:
                               Estimate Std. Error t value Pr(>|t|)    
(Intercept)                      78.407     27.717   2.829  0.00759 ** 
temperature                      37.481      1.704  21.995  < 2e-16 ***
locationOosterpark               99.644     42.527   2.343  0.02477 *  
temperature:locationOosterpark  -11.194      2.255  -4.963 1.68e-05 ***
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 36.86 on 36 degrees of freedom
Multiple R-squared:  0.9588,	Adjusted R-squared:  0.9554 
F-statistic: 279.6 on 3 and 36 DF,  p-value: < 2.2e-16

lgrdata documentation built on June 19, 2019, 5:03 p.m.