StatCrunch: Computing Correlations between Columns Video Lecture | Statcrunch Tutorials for Beginners - Data & Analytics

in this video you will learn to compute
the correlation between columns using
statcrunch the data set I will be using
is called home prices in Albuquerque
this dataset contains eight columns of
data taken from 117 home sales in
Albuquerque New Mexico in 1993 the price
column represents the sale price and
hundreds of dollars for each home so for
example the first value of 2050
represents a sales price of two hundred
and five thousand dollars the square
footage column represents the square
footage of the living space and the age
column is the age in years for each home
the feeds column gives the number of
features such as a dishwasher microwave
or dryer included in the home the next
three columns are indicator variables
with a value of one if the home has the
characteristic and a value of zero if
the home does not the any column
represents that the home is in the
northeast sector of the city the cust
column is if the home is a custom home
and the core column is if the home is on
a corner lot the tax column represents
the annual property taxes for the house
and dollars let's begin by creating a
correlation matrix between the variables
and the data set to do so under the stat
menu choose summary stats correlation
for this example I want to select all
the columns to do so hold down command
on a Mac or ctrl on a PC and choose all
of the columns now click compute and the
result is a correlation matrix where
each cell contains the sample
correlation between two variables so for
example in the cell between price and
square footage an approximate value of
0.8 4 represents the correlation between
the sales prices of the home and its
square footage
notice by default duplicate correlations
and correlations between a variable in
itself are left out of the matrix now
let's add p-values to each sample
correlation to help identify where there
is a significant association between
variables to do so under options choose
edit this takes us back to the original
window where we set up our matrix here
statcrunch has other options for the
user to choose under display I'm going
to choose two-sided p-value I'll click
compute the resulting correlation matrix
now has an additional line per cell that
contains the two-sided p-value for the
correlation between the corresponding
variables instead of all this extra
information let's see
say you just want to see a matrix how
price compares to the other variables
under options edit under display columns
I do not want to see all the columns I
only want to choose the price column so
I'll highlight price and click compute
now we have an altered correlation
matrix that displays only one column
showing the correlation of price with
the other characteristics however notice
that this is ordered by the columns and
the data set let's say you want to
change the ordering of the correlations
back under options choose edit I want to
order the columns by price and I can
choose ascending or descending for this
example I'll choose ascending and click
compute now the correlation matrix has
the variables across the rows sorted in
ascending order from the lowest
correlation price and age to the
correlation price with tax