I’m having trouble opening up that Powerpoint, but basically, in terms of a database…
When you normalize you store less data, and you don’t have gaps of empty columns.
Take a person for example
ID, First Name, Last Name, Date of Birth, Home Phone, Cell Phone, Work Phone, Fax Number
Not every person has a home, work, cell or fax number.
So to normalize it… you need 3 tables.
Members (ID, First Name, Last Name, Date of Birth)
Phone_Type (Phone_Type_ID, Description)
Members_Phone (Id, Phone_Type_ID, Number)
Phone_Type will have 4 rows
1, Home
2, Work
3, Cell
4, Fax
Members will have a lot of rows…
1, John, Doe, 1/1/1900
2, Jane, Doe, 1/1/1900
…
Members_Phone
1,1, 555-5555
1,2, 666-6666
2,1, 777-7777
…
This way you store the least amount of data possible, with no empty columns. It doesn’t make since to separate out Date of Birth, Last Name or First Name (because people usually only have 1 DOB, First Name and Last Name). But if you were going to capture (DOB, Wedding Date, Graduation Date, etc) then it will make sense to make a Special_Ocassion_Date_Types table and a Members_Special_Occassion Table.
So build a big table like you did with 71 columns, call it 1NF. Then start transforming it into 3NF.
Don’t worry about 4NF and 5NF (I hardly ever see that done in practice).
Then once you get table design down, you’ll work on adding in Foreign Keys to enforce integrity of your database. A lot of people will use an ER (Entity Relationship) Diagram tool to structure the database and make sure everything is connected visually and nothing is missed.