pairwise shifts data driven process improvement€¦ · benchmark test suite data driven process...

1
Benchmark Test Suite Data Driven Process Improvement Caroline Bonnes | Anthony D’Ambrosio | Jake Hansen | Robert Shelton Advisor: Steve Van Aken | Client: Jay Foster, Bob Majure Solution Architecture Next Steps Hypothesis Testing Flex-Metrics data are underutilized. throughput losses. Impact KPI Data Exploration Flex-Metrics (Flex) 700 production lines continue development additional use cases automation graphical user interface Learning Factory Client Background benchmark_stats(data, isSegmentData, area, equip, output_name) Production area of interest Pandas dataframe containing raw data Boolean for record type Equipment in area of interest Filename for excel output Opportunity Statement Estimated (Per Flex Client) $750,000 Minimum $700k Maximum $820k Automotive Supplier throughput overtime spending Input Aggregate results & Create visuals Tukey test for pairwise equipment Clean/Aggregate Tukey test for pairwise shifts Export Excel file. cleans, queries, and aggregates two Tukey tests. the best machine and shift what is happening in a production area. Increase Customer Throughput Data Source Excel Client Mfg. Plant Benchmark results and statistics (Excel) Input Z-Test comparing independent Variables Print Results run_hypothesis(‘TPut’, data, ‘Shift’, ‘1st Shift’, ‘2nd Shift’) Pandas dataframe containing raw data KPI of interest Independent variable group Independent variable instances answer questions using Z-Test. statistical differences in a KPI why machines are behaving a certain way Graphic view of spread Sample Function Call Sample Function Call

Upload: others

Post on 28-Jun-2020

2 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: pairwise shifts Data Driven Process Improvement€¦ · Benchmark Test Suite Data Driven Process Improvement Caroline Bonnes | Anthony D’Ambrosio | Jake Hansen | Robert Shelton

Benchmark Test Suite

Data Driven Process ImprovementCaroline Bonnes | Anthony D’Ambrosio | Jake Hansen | Robert Shelton

Advisor: Steve Van Aken | Client: Jay Foster, Bob Majure

Solution Architecture

Next Steps

Hypothesis Testing

Flex-Metrics data are underutilized.

throughput losses. Impact

KPI Data Exploration

Flex-Metrics (Flex)

700 production lines

continue development

● additional use cases

● automation● graphical user interface

● Learning Factory

Client Background

benchmark_stats(data, isSegmentData, area, equip, output_name)

Production area of interest

Pandas dataframe containing raw data

Boolean for record type

Equipment in area of interest

Filename for excel output

Opportunity Statement

Estimated(Per Flex Client)

$750,000Minimum$700k

Maximum$820k

Automotive Supplier

throughput

overtime spending

Input

Aggregate results & Create visuals

Tukey test for pairwise equipmentClean/Aggregate

Tukey test for pairwise shifts

Export Excel file.

cleans, queries, and aggregatestwo Tukey tests. the best machine and shift

what is happening in a production area.

IncreaseCustomer

Throughput

Data Source

Excel Client Mfg. Plant

Benchmark results and statistics(Excel)

InputZ-Test comparing

independent Variables

Print Results

run_hypothesis(‘TPut’, data, ‘Shift’, ‘1st Shift’, ‘2nd Shift’)

Pandas dataframe containing raw data

KPI of interest

Independent variable group

Independent variable instances

answer questions using Z-Test.statistical differences in a KPI

why machines are behaving a certain way

Graphic view of spread

Sample Function Call

Sample Function Call