14 Home

pt edited this page 2020-10-27 23:42:27 -04:00

Table of Contents

• Setup
• issues for consideration later

Start with a single table of sales data and be able to forecast it out. Also useful for simply reviewing sales data.

• Identify the table
• Label the columns of the table to generate master data tables
• Build a forecase baseline
• Start Making Changes

Setup

data prep check-list

• is master revisions buried in history (a part# could have 2 descriptions)
• open orders may not have a ship date, how would this be handled for baselining into next year
• ship dates may not reflect the fiscal period they are attributed to, should be snapped
• if the target table doesn't have version and iter they will need to be added
• sales history reflects only what intends to be forecast (gross sales, no canceled orders, short-ships)

issues for consideration later

• pull-forward is not well identified. If a customer already ordered for the season, shouldn't forecast it again.