UnsplashのAlexander Cifuentesが撮影した写真
This post is following of above post.
In this post, I will do cross sectional regression analysis using infer package.
Fisrt, I select one year for cross sectional regression. I check which year has the most number of observations.
2017 year has the most observations.
So, I make a data frame which contains only 2017.
Let's see summary statistics of y2017.
Let's see scatter plots.
Then, I load infer package.
In this post, I will analysis per_capita ~ acc_nights + inter_arr + inter_dep model with traditional lm() function eco system and infer package eco system.
Fisrt, I make the both regression results.
the both results gets the same slope estimates.
acc_nights is 0.142,
inter_arr is -0.605
inter_dep is 2.62.
How about 95% confidence intervals?
To get confidence intervals with infer ecosystem, I need to make null distribution of slope estimates.
Then, I use get_confidence_interval() function.
In lm() ecosystem, I can use confint() function to get confidence intervals.
All right, let's visualize confidence intervals.
First, acc_nights confidence intervals
The both confident intervals include zero, so, acc_nights is not statistically significant.
Next, inter_arr.
Again, the both confidence intervals include zero, so that inter_arr is not statistically significant.
Last, inter_dep.
The both inter_dep slope estimate confidence intervals contain zero, so inter_dep is not statistically significant.
That's it. Thank you!
To read from the 1st post,
