This document outlines an approach using CRISP-DM (Cross Industry Standard Process for Data Mining) to analyze food procurement prices across a supply chain. The goal is to determine if prices paid are optimal by identifying outliers, comparing categories and channels. Key steps include understanding the industry and data, preparing data by standardizing units of measure, modeling prices statistically, evaluating outliers and presenting findings to identify opportunities to negotiate better prices.
4. Are we buying right item at right price uniformly?
Statistical Question
5. Approach
Cross Industry Standard Process for Data Mining (CRISP-DM)
Business Understanding
Data Understanding
Data Preparation
Modeling
Evaluation
Deployment
6. CRISP-DM : Business Understanding
Food Industry Features
Many substitutes available
Same item from different manufacturer
De-Centralize buying
Small local vendor like bakery
Supplier base delivery
Freight/Order size
Multi channel sourcing manufacturer item/supplier item 1:n
Number of items
Weekly price publication
7. CRISP-DM : Data Understanding
Prices depends upon
Quantity
Item Size
Item weight
Pack size varies by quantity
Variety of unit of measure
Items can be classified into groups
Meat, Dairy, Produce,
Dairy Milk, Cheese, Yoghurt,
Milk Fat Free, 2% Fat Free,
8. CRISP-DM : Data Preparation
Collect catalog price
Identify prices with
Ingredient
Business segment
Determine Uniform Unit of Measure
Calculate unit level price
9. CRISP-DM : Modelling
Analysis Grouping
Ingredient
Item Classification
Manufacturer Item
Statistical Analysis
Data Analysis
Determine central values
Mean
Median
Determine Spread
Range
Standard Deviation
Z- Score
Data presentation
Plot
10. CRISP-DM : Evaluation
Determine data properties
Outliers
Range
Anomaly
Let data tell its story
Graphs
Excel