I'm currently running some rudimentary exploratory data analysis on the common Kaggle Titanic survival rate competition, and would like to know how to interpret the following heatmap correlation plot I have made below.
I understand positive correlation (relating to pearsons r-coefficient) - in that as X increases Y also increases and the closer to +1 indicating a strong positive linear relationship.
But for negative correlation values I am quite perplexed as from my understanding (which may be wrong please correct me if so) that the relationship is inverted - so as X increases Y tends to decrease implying if a value has close to -1 then that is a strong negative linear relationship.
I'm struggling with how to interpret this from the heatmap below. The feature that contains the binary outcome I am trying to predict is the feature Survived The rest of the columns are numeric columns that are used for the prediction.
Here we can see that Age/Survived = -0.06 and SibSp/Survived = -0.03
How do i interpret this? Am i correct in saying the following:
- As Age Decreases, Survival Rate Increases (weak negative correlation)
- As SibSp Decreases, Survival Rate Increases (weak negative correlation)
As a side note if the values were -0.90 so close to -1 - how do I word the explanation correctly, would it be as for example: Age decreases then Survival rate increases?