A Survey on Retail Sales - 20th July, 2013

In today's lecture we worked on retail stores data. We learned how to dig deep into data and come up with relations between various variables. Which variables are related to each other and which are not.

So here is the variable view of the data.

First we tried to find out which variables are nominal and which are ordinal. As we already know :-

Nominal -  It is the lowest level of data measurement . These numbers don't have any meaning . These can only be used to classify or categorize .

Ordinal - It is the second level of data measurement . These numbers can be used to rank or order objects.

Then we analysed the frequency for age category variable to see the percentage of people of various age groups.

Then we learned a new concept of crosstab.

Crosstab - In statistics, a "crosstab" is another name for a contingency table, which is a type of table created by crosstabulation. In survey research (e.g., polling, market research), a "crosstab" is any table showing summary statistics. Commonly, crosstabs in survey research are concatenations of multiple different tables.

We did cross tab for Store and Service satisfaction variables.

This is how we do crosstab.

Analyse -> Descriptive Statistics -> Crosstabs

Here we put store in rows as we are comparing for stores.

For store 1 as we can see from the table 17.1% of the total people who visited store 1 are stronly negetive and 26.9% of the total people who are strongly negetive are negetive for store 1.

Next we learned about Null Hypothesis and CHI-Square

The null hypothesis in cross tab says that there is no relationship between the two variables we are testing.

Significance value - < 0.05 reject

                              > 0.05 accept

When it is more than 0.05 we accept it and it means there is no relation between the two variables.

Then we did the correlation of various satisfactions.

If the correlation value is high then if one variable is high than other correlation variable is also high.

When the numbers get low, stop analysis.

For any deep analysis take big sample space.

With this we ended the lecture.

Blog written by :-

 Piyush Mittal - 2013197

Group Members

   Prerna Bansal

   Priya Jain

   Neeraj Garg

   Piyush Mittal

   Pallavi Gupta