Please enable JavaScript.
Coggle requires JavaScript to display documents.
EDA (Exploring Numerical Data (Measure of centrality (Median better for…
EDA
Exploring Numerical Data
-
- Using boxplots to understand median/s, IQRs, outlies
- Can check associations of one var by another using faceting. Keep an eye for variability and typical obs
- Can find distr of a specific occurence using xlim() or filter combined with a hist
- Can do multi-variate analysis by faceting via multiple variables. If using smthin like a histogram, check contingency table to check if obs not negligible.
-
-
Can be unimodal, bimodal, multi-modal or uniform
Right, left or symmetrical
Logging or sq rooting reduces some skew for better shape analysis (rmr that can reverse calc as well). Also, rmr -inf case
-
-