SlideShare a Scribd company logo
EDBT 2021
Assess Queries for Interactive
Analysis of Data Cubes
Matteo Francia1, Matteo Golfarelli1, Patrick Marcel2, Stefano Rizzi1, Panos Vassiliadis3
1University of Bologna, Italy; 2University of Tours, France; 3University of Ioannina, Greece
24th International Conference on Extending Database Technology (EDBT 2021)
EDBT 2021
Motivation
Goal: compare business phenomena over multidimensional cubes
OLAP provides (low-level) operators to query cubes [1]
- Need higher abstractions to express analytic intentions
Some operators complement the fundamental ones [2]
- E.g., Descriptive [3] or Predictive [4] purposes
We introduce the Assess operator
- Compare the actual to the expected behavior
- Judge the outcome of the comparison
Matteo Francia – University of Bologna 2
Introduction
[1] Panos Vassiliadis, Patrick Marcel, Stefano Rizzi: Beyond roll-up's and drill-down's: An intentional analytics model to reinvent OLAP. Inf. Syst. 85: 68-91 (2019)
[2] Lukasz Golab, Divesh Srivastava: Exploring Data Using Patterns: A Survey and Open Problems. DOLAP@EDBT2021
[3] Sunita Sarawagi: Explaining Differences in Multidimensional Aggregates. VLDB 1999: 42-53
[4] Bee-Chung Chen, Lei Chen, Yi Lin, Raghu Ramakrishnan: Prediction Cubes. VLDB 2005: 982-993
EDBT 2021
… in classical OLAP vs IAM
Intentional Analytics Model (IAM) [1,2]
- Bring OLAP to higher abstraction level
- Escape from query answers as sets of tuples
Express intentions, not queries
- High level syntax
- Semantics
- Logical/physical optimization
with cube [ for predicate ] by levels assess measure
[ against <benchmark> ] [ using <function> ]
labels 𝜆
Matteo Francia – University of Bologna
[1] Panos Vassiliadis, Patrick Marcel, Stefano Rizzi: Beyond roll-up's and drill-down's: An intentional analytics model to reinvent OLAP. Inf. Syst. 85: 68-91 (2019)
[2] Antoine Chédin, Matteo Francia, Patrick Marcel, Verónika Peralta, Stefano Rizzi: The Tell-Tale Cube. ADBIS 2020: 204-218
Query the cube
Manual comparison
- Hard if transformations are needed
- Hard with multiple cubes
select year, product, sum(quantity)
from sales_ft join date_dt on (…)
where year = 2019
group by year, product
year product sum(quantity)
2019 Milk 125
2019 Apple 100
… … …
Introduction
EDBT 2021
Computing an Assessment
Assess a measure 𝑚 in a cube 𝐶 (target cube)
Specify a benchmark cube 𝐵 that
- Has a measure 𝑚′ representing the expected behavior of 𝑚
- Is joinable with 𝐶 (𝐵 and 𝐶 might coincide)
Optionally apply transformations
- E.g., normalization
Compare 𝑚 to 𝑚′ cell-wise
- E.g., difference and ratio
Label cells of 𝐶 based on the comparison
- Map the result of the comparison to predefined labels
Matteo Francia – University of Bologna 4
𝐶 𝐵
𝑚: quantity = 100 𝑚′: sales = 80
𝐶
C.quantity – B.sales > 0
Assess
Operator
EDBT 2021
Computing an Assessment
4 benchmark types
- Fixed: compare a cube against a fixed value
- Sibling: compare two slices from a cube
- Past: compare a slice against a historic-based expectation
- External: compare two cubes
Compare siblings in 𝐶
- E.g., Italian sales are good if higher than French sales
with SALES for country = ‘Italy’ by product, country
assess quantity against country = ‘France’
using percOfTotal(difference(quantity, benchmark.quantity))
labels {[-inf, -0.2): bad, [-0.2,0.2]: ok, (0.2, inf ]: good}
Matteo Francia – University of Bologna 5
Assess
Operator
SALES
Apple
Pear
Lemon
Cola
Pizza
Beer
EDBT 2021
Implementing Assess: 5 Operators
Get: given 𝐶, predicate 𝑃, and group-by set 𝐺, returns 𝐶’
Cell-Transform ⊟: cell-at-a-time transformation
- Returns 𝐶’ with a new measure value for each cell
H-Transform ⊡: holistic transformation
- Returns 𝐶’ with a new measure value for each cell
given all cells of 𝐶
Matteo Francia – University of Bologna 6
𝐶 𝐶’
𝐶 𝐶’
𝐶
𝐶’
Implementation
EDBT 2021
Implementing Assess: 5 Operators
(Partial) Join ⊠: put together cube and benchmark
Pivot ⊞: pivot member’s measure values to new measures
Matteo Francia – University of Bologna 7
=
⊠
product
⊞
country,
country=‘Italy’
Implementation
SALES
Lemon
quantity=100
COMP
Lemon
quantity=80
SALES
Lemon
quantity=100
C’
Lemon
quantity=100
quantityFrance=80
=
C’
Lemon
quantity=100
quantityFrance=80
EDBT 2021
SALES
Apple
Pear
Lemon
Cola
Pizza
Beer
Sibling Benchmark: Semantics
with SALES for country = ‘Italy’ by product, country
assess quantity against country = ‘France’
using percOfTotal(difference(quantity, benchmark.quantity))
labels {[-inf, -0.2): bad, [-0.2,0.2]: ok, (0.2, inf ]: good}
Get 𝐶 and 𝐵
Partial join on product
Transform & label
Matteo Francia – University of Bologna 8
Execution
Plan
Apple
Pear
Lemon
Cola
Pizza
Beer
quantity = 100
quantityFrance = 80
Apple
Pear
Lemon
Cola
Pizza
Beer
quantity=100, quantityFrance=80
diff=20, perc=0.4,label=good
EDBT 2021
Can we do better?
Intuition: push operators into DBMS to exploit optimization structures
From NP to JOP: push cell-wise transformations after join
From JOP to POP: transform Join into Pivot
Matteo Francia – University of Bologna 9
Naive Plan (NP) Join Optimized Plan (JOP) Pivot Optimized Plan (POP)
1. Get C (SQL)
2. Get B (SQL)
3. Join C and B (memory)
4. Transform (memory)
5. Label (memory)
1. Join C and B (SQL)
2. Transform (memory)
3. Label (memory)
1. Pivot slices in C (SQL)
2. Transform (memory)
3. Label (memory)
Execution
Plan
EDBT 2021
Cube cardinality: SSBi ~= 6𝑀 ⋅ 𝑖 tuples
All plans scale linearly
Alternative execution plans
- JOP faster than NP (when applicable)
- POP faster than JOP (when applicable)
Past is the heaviest benchmark due to trend estimation
Empirical Evaluation
Matteo Francia – University of Bologna 10
Evaluation
&
Conclusion
EDBT 2021
Conclusion
The assess operator: evaluate cube against benchmark, label the result
- Several alternatives for benchmarks, comparison, and labeling schema
Alternative execution plans, increase performance by pushing to DBMS
Future work
- Devise strategies for effectively completing partial assess statements
- E.g., if no against is specified, infer the most interesting benchmark
- Enhance the expressiveness of the assess
- E.g., labeling based on ranges that depend on the coordinates
- Investigate operators to develop a cost-based optimization strategy
Matteo Francia – University of Bologna 11
Evaluation
&
Conclusion
EDBT 2021
Questions?
Matteo Francia – University of Bologna 12
Thank you. Compute query
& benchmark
Map cells &
compare
Rule-based cell
labeling
SALES
http://www.cs.uoi.gr/~pvassil/projects/olap_III/

More Related Content

Similar to [EDBT2021] Assess Queries for Interactive Analysis of Data Cubes

Microsoft Office Performance Point
Microsoft Office Performance PointMicrosoft Office Performance Point
Microsoft Office Performance Point
Kashif Akram
 
Online Intelligence- POL
Online Intelligence- POLOnline Intelligence- POL
Online Intelligence- POLDavid Barak
 
[EDBT2023] Describing and Assessing Cubes Through Intentional Analytics (demo...
[EDBT2023] Describing and Assessing Cubes Through Intentional Analytics (demo...[EDBT2023] Describing and Assessing Cubes Through Intentional Analytics (demo...
[EDBT2023] Describing and Assessing Cubes Through Intentional Analytics (demo...
University of Bologna
 
WELCOMEFinancial Projections ModelFor Business PlansFran.docx
WELCOMEFinancial Projections ModelFor Business PlansFran.docxWELCOMEFinancial Projections ModelFor Business PlansFran.docx
WELCOMEFinancial Projections ModelFor Business PlansFran.docx
alanfhall8953
 
WELCOMEFinancial Projections ModelFor Business PlansFran.docx
WELCOMEFinancial Projections ModelFor Business PlansFran.docxWELCOMEFinancial Projections ModelFor Business PlansFran.docx
WELCOMEFinancial Projections ModelFor Business PlansFran.docx
philipnelson29183
 
Advanced Optimization for the Enterprise Webinar
Advanced Optimization for the Enterprise WebinarAdvanced Optimization for the Enterprise Webinar
Advanced Optimization for the Enterprise Webinar
SigOpt
 
How can Digital Twins support Manufacturers on the path to Net-Zero?
How can Digital Twins support Manufacturers on the path to Net-Zero?How can Digital Twins support Manufacturers on the path to Net-Zero?
How can Digital Twins support Manufacturers on the path to Net-Zero?
IES VE
 
GBS Benchmarking Solutions
GBS Benchmarking SolutionsGBS Benchmarking Solutions
GBS Benchmarking Solutions
Philippe FIXEL
 
Tutorial_Caf_2020_etool.pdf
Tutorial_Caf_2020_etool.pdfTutorial_Caf_2020_etool.pdf
Tutorial_Caf_2020_etool.pdf
Salvatore Spampinato
 
5 Steps To Measure ROI On Your Data Science Initiatives - Webinar
 5 Steps To Measure ROI On Your Data Science Initiatives - Webinar 5 Steps To Measure ROI On Your Data Science Initiatives - Webinar
5 Steps To Measure ROI On Your Data Science Initiatives - Webinar
Gramener
 
IENG 302 Lecture 11_Benefit cost analysis.ppt
IENG 302 Lecture 11_Benefit cost analysis.pptIENG 302 Lecture 11_Benefit cost analysis.ppt
IENG 302 Lecture 11_Benefit cost analysis.ppt
VitaliyBondar4
 
Introduction to Multiple Regression
Introduction to Multiple RegressionIntroduction to Multiple Regression
Introduction to Multiple Regression
Yesica Adicondro
 
2011.06.23 How to Value Software in a Business
2011.06.23 How to Value Software in a Business2011.06.23 How to Value Software in a Business
2011.06.23 How to Value Software in a Business
NUI Galway
 
TabPy Presentation
TabPy PresentationTabPy Presentation
TabPy Presentation
Sanjana Jami
 
Applied Machine Learning for Ranking Products in an Ecommerce Setting
Applied Machine Learning for Ranking Products in an Ecommerce SettingApplied Machine Learning for Ranking Products in an Ecommerce Setting
Applied Machine Learning for Ranking Products in an Ecommerce Setting
Databricks
 
Vision 2025 du groupe Epson
Vision 2025 du groupe EpsonVision 2025 du groupe Epson
Vision 2025 du groupe Epson
Jean-Marc BOURNE
 
1 ISM645 Strategic Information Technology Planning v1.1 .docx
1 ISM645 Strategic Information Technology Planning v1.1 .docx1 ISM645 Strategic Information Technology Planning v1.1 .docx
1 ISM645 Strategic Information Technology Planning v1.1 .docx
teresehearn
 
1 ISM645 Strategic Information Technology Planning v1.1 .docx
1 ISM645 Strategic Information Technology Planning v1.1 .docx1 ISM645 Strategic Information Technology Planning v1.1 .docx
1 ISM645 Strategic Information Technology Planning v1.1 .docx
karisariddell
 
Data Science for Smart Manufacturing
Data Science for Smart ManufacturingData Science for Smart Manufacturing
Data Science for Smart Manufacturing
Carlo Torniai
 
Predictive Data Analytics and Artificial Intelligence by 40°
Predictive Data Analytics and Artificial Intelligence by 40°Predictive Data Analytics and Artificial Intelligence by 40°
Predictive Data Analytics and Artificial Intelligence by 40°
40° Labor für Innovation
 

Similar to [EDBT2021] Assess Queries for Interactive Analysis of Data Cubes (20)

Microsoft Office Performance Point
Microsoft Office Performance PointMicrosoft Office Performance Point
Microsoft Office Performance Point
 
Online Intelligence- POL
Online Intelligence- POLOnline Intelligence- POL
Online Intelligence- POL
 
[EDBT2023] Describing and Assessing Cubes Through Intentional Analytics (demo...
[EDBT2023] Describing and Assessing Cubes Through Intentional Analytics (demo...[EDBT2023] Describing and Assessing Cubes Through Intentional Analytics (demo...
[EDBT2023] Describing and Assessing Cubes Through Intentional Analytics (demo...
 
WELCOMEFinancial Projections ModelFor Business PlansFran.docx
WELCOMEFinancial Projections ModelFor Business PlansFran.docxWELCOMEFinancial Projections ModelFor Business PlansFran.docx
WELCOMEFinancial Projections ModelFor Business PlansFran.docx
 
WELCOMEFinancial Projections ModelFor Business PlansFran.docx
WELCOMEFinancial Projections ModelFor Business PlansFran.docxWELCOMEFinancial Projections ModelFor Business PlansFran.docx
WELCOMEFinancial Projections ModelFor Business PlansFran.docx
 
Advanced Optimization for the Enterprise Webinar
Advanced Optimization for the Enterprise WebinarAdvanced Optimization for the Enterprise Webinar
Advanced Optimization for the Enterprise Webinar
 
How can Digital Twins support Manufacturers on the path to Net-Zero?
How can Digital Twins support Manufacturers on the path to Net-Zero?How can Digital Twins support Manufacturers on the path to Net-Zero?
How can Digital Twins support Manufacturers on the path to Net-Zero?
 
GBS Benchmarking Solutions
GBS Benchmarking SolutionsGBS Benchmarking Solutions
GBS Benchmarking Solutions
 
Tutorial_Caf_2020_etool.pdf
Tutorial_Caf_2020_etool.pdfTutorial_Caf_2020_etool.pdf
Tutorial_Caf_2020_etool.pdf
 
5 Steps To Measure ROI On Your Data Science Initiatives - Webinar
 5 Steps To Measure ROI On Your Data Science Initiatives - Webinar 5 Steps To Measure ROI On Your Data Science Initiatives - Webinar
5 Steps To Measure ROI On Your Data Science Initiatives - Webinar
 
IENG 302 Lecture 11_Benefit cost analysis.ppt
IENG 302 Lecture 11_Benefit cost analysis.pptIENG 302 Lecture 11_Benefit cost analysis.ppt
IENG 302 Lecture 11_Benefit cost analysis.ppt
 
Introduction to Multiple Regression
Introduction to Multiple RegressionIntroduction to Multiple Regression
Introduction to Multiple Regression
 
2011.06.23 How to Value Software in a Business
2011.06.23 How to Value Software in a Business2011.06.23 How to Value Software in a Business
2011.06.23 How to Value Software in a Business
 
TabPy Presentation
TabPy PresentationTabPy Presentation
TabPy Presentation
 
Applied Machine Learning for Ranking Products in an Ecommerce Setting
Applied Machine Learning for Ranking Products in an Ecommerce SettingApplied Machine Learning for Ranking Products in an Ecommerce Setting
Applied Machine Learning for Ranking Products in an Ecommerce Setting
 
Vision 2025 du groupe Epson
Vision 2025 du groupe EpsonVision 2025 du groupe Epson
Vision 2025 du groupe Epson
 
1 ISM645 Strategic Information Technology Planning v1.1 .docx
1 ISM645 Strategic Information Technology Planning v1.1 .docx1 ISM645 Strategic Information Technology Planning v1.1 .docx
1 ISM645 Strategic Information Technology Planning v1.1 .docx
 
1 ISM645 Strategic Information Technology Planning v1.1 .docx
1 ISM645 Strategic Information Technology Planning v1.1 .docx1 ISM645 Strategic Information Technology Planning v1.1 .docx
1 ISM645 Strategic Information Technology Planning v1.1 .docx
 
Data Science for Smart Manufacturing
Data Science for Smart ManufacturingData Science for Smart Manufacturing
Data Science for Smart Manufacturing
 
Predictive Data Analytics and Artificial Intelligence by 40°
Predictive Data Analytics and Artificial Intelligence by 40°Predictive Data Analytics and Artificial Intelligence by 40°
Predictive Data Analytics and Artificial Intelligence by 40°
 

More from University of Bologna

Data models in precision agriculture: from IoT to big data analytics
Data models in precision agriculture: from IoT to big data analyticsData models in precision agriculture: from IoT to big data analytics
Data models in precision agriculture: from IoT to big data analytics
University of Bologna
 
[DataPlat2023] Opening
[DataPlat2023] Opening[DataPlat2023] Opening
[DataPlat2023] Opening
University of Bologna
 
[ADBIS2022] Insight-based Vocalization of OLAP Sessions
[ADBIS2022] Insight-based Vocalization of OLAP Sessions[ADBIS2022] Insight-based Vocalization of OLAP Sessions
[ADBIS2022] Insight-based Vocalization of OLAP Sessions
University of Bologna
 
[PhDThesis2021] - Augmenting the knowledge pyramid with unconventional data a...
[PhDThesis2021] - Augmenting the knowledge pyramid with unconventional data a...[PhDThesis2021] - Augmenting the knowledge pyramid with unconventional data a...
[PhDThesis2021] - Augmenting the knowledge pyramid with unconventional data a...
University of Bologna
 
[SEBD2020] OLAP Querying of Document Stores in the Presence of Schema Variety
[SEBD2020] OLAP Querying of Document Stores in the Presence of Schema Variety[SEBD2020] OLAP Querying of Document Stores in the Presence of Schema Variety
[SEBD2020] OLAP Querying of Document Stores in the Presence of Schema Variety
University of Bologna
 
[MIPRO2019] Map-Matching on Big Data: a Distributed and Efficient Algorithm w...
[MIPRO2019] Map-Matching on Big Data: a Distributed and Efficient Algorithm w...[MIPRO2019] Map-Matching on Big Data: a Distributed and Efficient Algorithm w...
[MIPRO2019] Map-Matching on Big Data: a Distributed and Efficient Algorithm w...
University of Bologna
 
[DOLAP2019] Augmented Business Intelligence
[DOLAP2019] Augmented Business Intelligence[DOLAP2019] Augmented Business Intelligence
[DOLAP2019] Augmented Business Intelligence
University of Bologna
 

More from University of Bologna (7)

Data models in precision agriculture: from IoT to big data analytics
Data models in precision agriculture: from IoT to big data analyticsData models in precision agriculture: from IoT to big data analytics
Data models in precision agriculture: from IoT to big data analytics
 
[DataPlat2023] Opening
[DataPlat2023] Opening[DataPlat2023] Opening
[DataPlat2023] Opening
 
[ADBIS2022] Insight-based Vocalization of OLAP Sessions
[ADBIS2022] Insight-based Vocalization of OLAP Sessions[ADBIS2022] Insight-based Vocalization of OLAP Sessions
[ADBIS2022] Insight-based Vocalization of OLAP Sessions
 
[PhDThesis2021] - Augmenting the knowledge pyramid with unconventional data a...
[PhDThesis2021] - Augmenting the knowledge pyramid with unconventional data a...[PhDThesis2021] - Augmenting the knowledge pyramid with unconventional data a...
[PhDThesis2021] - Augmenting the knowledge pyramid with unconventional data a...
 
[SEBD2020] OLAP Querying of Document Stores in the Presence of Schema Variety
[SEBD2020] OLAP Querying of Document Stores in the Presence of Schema Variety[SEBD2020] OLAP Querying of Document Stores in the Presence of Schema Variety
[SEBD2020] OLAP Querying of Document Stores in the Presence of Schema Variety
 
[MIPRO2019] Map-Matching on Big Data: a Distributed and Efficient Algorithm w...
[MIPRO2019] Map-Matching on Big Data: a Distributed and Efficient Algorithm w...[MIPRO2019] Map-Matching on Big Data: a Distributed and Efficient Algorithm w...
[MIPRO2019] Map-Matching on Big Data: a Distributed and Efficient Algorithm w...
 
[DOLAP2019] Augmented Business Intelligence
[DOLAP2019] Augmented Business Intelligence[DOLAP2019] Augmented Business Intelligence
[DOLAP2019] Augmented Business Intelligence
 

Recently uploaded

Q1’2024 Update: MYCI’s Leap Year Rebound
Q1’2024 Update: MYCI’s Leap Year ReboundQ1’2024 Update: MYCI’s Leap Year Rebound
Q1’2024 Update: MYCI’s Leap Year Rebound
Oppotus
 
一比一原版(BCU毕业证书)伯明翰城市大学毕业证如何办理
一比一原版(BCU毕业证书)伯明翰城市大学毕业证如何办理一比一原版(BCU毕业证书)伯明翰城市大学毕业证如何办理
一比一原版(BCU毕业证书)伯明翰城市大学毕业证如何办理
dwreak4tg
 
Algorithmic optimizations for Dynamic Levelwise PageRank (from STICD) : SHORT...
Algorithmic optimizations for Dynamic Levelwise PageRank (from STICD) : SHORT...Algorithmic optimizations for Dynamic Levelwise PageRank (from STICD) : SHORT...
Algorithmic optimizations for Dynamic Levelwise PageRank (from STICD) : SHORT...
Subhajit Sahu
 
Malana- Gimlet Market Analysis (Portfolio 2)
Malana- Gimlet Market Analysis (Portfolio 2)Malana- Gimlet Market Analysis (Portfolio 2)
Malana- Gimlet Market Analysis (Portfolio 2)
TravisMalana
 
一比一原版(Deakin毕业证书)迪肯大学毕业证如何办理
一比一原版(Deakin毕业证书)迪肯大学毕业证如何办理一比一原版(Deakin毕业证书)迪肯大学毕业证如何办理
一比一原版(Deakin毕业证书)迪肯大学毕业证如何办理
oz8q3jxlp
 
原版制作(Deakin毕业证书)迪肯大学毕业证学位证一模一样
原版制作(Deakin毕业证书)迪肯大学毕业证学位证一模一样原版制作(Deakin毕业证书)迪肯大学毕业证学位证一模一样
原版制作(Deakin毕业证书)迪肯大学毕业证学位证一模一样
u86oixdj
 
Criminal IP - Threat Hunting Webinar.pdf
Criminal IP - Threat Hunting Webinar.pdfCriminal IP - Threat Hunting Webinar.pdf
Criminal IP - Threat Hunting Webinar.pdf
Criminal IP
 
一比一原版(爱大毕业证书)爱丁堡大学毕业证如何办理
一比一原版(爱大毕业证书)爱丁堡大学毕业证如何办理一比一原版(爱大毕业证书)爱丁堡大学毕业证如何办理
一比一原版(爱大毕业证书)爱丁堡大学毕业证如何办理
g4dpvqap0
 
Everything you wanted to know about LIHTC
Everything you wanted to know about LIHTCEverything you wanted to know about LIHTC
Everything you wanted to know about LIHTC
Roger Valdez
 
Data_and_Analytics_Essentials_Architect_an_Analytics_Platform.pptx
Data_and_Analytics_Essentials_Architect_an_Analytics_Platform.pptxData_and_Analytics_Essentials_Architect_an_Analytics_Platform.pptx
Data_and_Analytics_Essentials_Architect_an_Analytics_Platform.pptx
AnirbanRoy608946
 
Influence of Marketing Strategy and Market Competition on Business Plan
Influence of Marketing Strategy and Market Competition on Business PlanInfluence of Marketing Strategy and Market Competition on Business Plan
Influence of Marketing Strategy and Market Competition on Business Plan
jerlynmaetalle
 
一比一原版(CBU毕业证)卡普顿大学毕业证如何办理
一比一原版(CBU毕业证)卡普顿大学毕业证如何办理一比一原版(CBU毕业证)卡普顿大学毕业证如何办理
一比一原版(CBU毕业证)卡普顿大学毕业证如何办理
ahzuo
 
My burning issue is homelessness K.C.M.O.
My burning issue is homelessness K.C.M.O.My burning issue is homelessness K.C.M.O.
My burning issue is homelessness K.C.M.O.
rwarrenll
 
一比一原版(UniSA毕业证书)南澳大学毕业证如何办理
一比一原版(UniSA毕业证书)南澳大学毕业证如何办理一比一原版(UniSA毕业证书)南澳大学毕业证如何办理
一比一原版(UniSA毕业证书)南澳大学毕业证如何办理
slg6lamcq
 
一比一原版(UPenn毕业证)宾夕法尼亚大学毕业证成绩单
一比一原版(UPenn毕业证)宾夕法尼亚大学毕业证成绩单一比一原版(UPenn毕业证)宾夕法尼亚大学毕业证成绩单
一比一原版(UPenn毕业证)宾夕法尼亚大学毕业证成绩单
ewymefz
 
Adjusting OpenMP PageRank : SHORT REPORT / NOTES
Adjusting OpenMP PageRank : SHORT REPORT / NOTESAdjusting OpenMP PageRank : SHORT REPORT / NOTES
Adjusting OpenMP PageRank : SHORT REPORT / NOTES
Subhajit Sahu
 
原版制作(swinburne毕业证书)斯威本科技大学毕业证毕业完成信一模一样
原版制作(swinburne毕业证书)斯威本科技大学毕业证毕业完成信一模一样原版制作(swinburne毕业证书)斯威本科技大学毕业证毕业完成信一模一样
原版制作(swinburne毕业证书)斯威本科技大学毕业证毕业完成信一模一样
u86oixdj
 
哪里卖(usq毕业证书)南昆士兰大学毕业证研究生文凭证书托福证书原版一模一样
哪里卖(usq毕业证书)南昆士兰大学毕业证研究生文凭证书托福证书原版一模一样哪里卖(usq毕业证书)南昆士兰大学毕业证研究生文凭证书托福证书原版一模一样
哪里卖(usq毕业证书)南昆士兰大学毕业证研究生文凭证书托福证书原版一模一样
axoqas
 
一比一原版(UofM毕业证)明尼苏达大学毕业证成绩单
一比一原版(UofM毕业证)明尼苏达大学毕业证成绩单一比一原版(UofM毕业证)明尼苏达大学毕业证成绩单
一比一原版(UofM毕业证)明尼苏达大学毕业证成绩单
ewymefz
 
一比一原版(Coventry毕业证书)考文垂大学毕业证如何办理
一比一原版(Coventry毕业证书)考文垂大学毕业证如何办理一比一原版(Coventry毕业证书)考文垂大学毕业证如何办理
一比一原版(Coventry毕业证书)考文垂大学毕业证如何办理
74nqk8xf
 

Recently uploaded (20)

Q1’2024 Update: MYCI’s Leap Year Rebound
Q1’2024 Update: MYCI’s Leap Year ReboundQ1’2024 Update: MYCI’s Leap Year Rebound
Q1’2024 Update: MYCI’s Leap Year Rebound
 
一比一原版(BCU毕业证书)伯明翰城市大学毕业证如何办理
一比一原版(BCU毕业证书)伯明翰城市大学毕业证如何办理一比一原版(BCU毕业证书)伯明翰城市大学毕业证如何办理
一比一原版(BCU毕业证书)伯明翰城市大学毕业证如何办理
 
Algorithmic optimizations for Dynamic Levelwise PageRank (from STICD) : SHORT...
Algorithmic optimizations for Dynamic Levelwise PageRank (from STICD) : SHORT...Algorithmic optimizations for Dynamic Levelwise PageRank (from STICD) : SHORT...
Algorithmic optimizations for Dynamic Levelwise PageRank (from STICD) : SHORT...
 
Malana- Gimlet Market Analysis (Portfolio 2)
Malana- Gimlet Market Analysis (Portfolio 2)Malana- Gimlet Market Analysis (Portfolio 2)
Malana- Gimlet Market Analysis (Portfolio 2)
 
一比一原版(Deakin毕业证书)迪肯大学毕业证如何办理
一比一原版(Deakin毕业证书)迪肯大学毕业证如何办理一比一原版(Deakin毕业证书)迪肯大学毕业证如何办理
一比一原版(Deakin毕业证书)迪肯大学毕业证如何办理
 
原版制作(Deakin毕业证书)迪肯大学毕业证学位证一模一样
原版制作(Deakin毕业证书)迪肯大学毕业证学位证一模一样原版制作(Deakin毕业证书)迪肯大学毕业证学位证一模一样
原版制作(Deakin毕业证书)迪肯大学毕业证学位证一模一样
 
Criminal IP - Threat Hunting Webinar.pdf
Criminal IP - Threat Hunting Webinar.pdfCriminal IP - Threat Hunting Webinar.pdf
Criminal IP - Threat Hunting Webinar.pdf
 
一比一原版(爱大毕业证书)爱丁堡大学毕业证如何办理
一比一原版(爱大毕业证书)爱丁堡大学毕业证如何办理一比一原版(爱大毕业证书)爱丁堡大学毕业证如何办理
一比一原版(爱大毕业证书)爱丁堡大学毕业证如何办理
 
Everything you wanted to know about LIHTC
Everything you wanted to know about LIHTCEverything you wanted to know about LIHTC
Everything you wanted to know about LIHTC
 
Data_and_Analytics_Essentials_Architect_an_Analytics_Platform.pptx
Data_and_Analytics_Essentials_Architect_an_Analytics_Platform.pptxData_and_Analytics_Essentials_Architect_an_Analytics_Platform.pptx
Data_and_Analytics_Essentials_Architect_an_Analytics_Platform.pptx
 
Influence of Marketing Strategy and Market Competition on Business Plan
Influence of Marketing Strategy and Market Competition on Business PlanInfluence of Marketing Strategy and Market Competition on Business Plan
Influence of Marketing Strategy and Market Competition on Business Plan
 
一比一原版(CBU毕业证)卡普顿大学毕业证如何办理
一比一原版(CBU毕业证)卡普顿大学毕业证如何办理一比一原版(CBU毕业证)卡普顿大学毕业证如何办理
一比一原版(CBU毕业证)卡普顿大学毕业证如何办理
 
My burning issue is homelessness K.C.M.O.
My burning issue is homelessness K.C.M.O.My burning issue is homelessness K.C.M.O.
My burning issue is homelessness K.C.M.O.
 
一比一原版(UniSA毕业证书)南澳大学毕业证如何办理
一比一原版(UniSA毕业证书)南澳大学毕业证如何办理一比一原版(UniSA毕业证书)南澳大学毕业证如何办理
一比一原版(UniSA毕业证书)南澳大学毕业证如何办理
 
一比一原版(UPenn毕业证)宾夕法尼亚大学毕业证成绩单
一比一原版(UPenn毕业证)宾夕法尼亚大学毕业证成绩单一比一原版(UPenn毕业证)宾夕法尼亚大学毕业证成绩单
一比一原版(UPenn毕业证)宾夕法尼亚大学毕业证成绩单
 
Adjusting OpenMP PageRank : SHORT REPORT / NOTES
Adjusting OpenMP PageRank : SHORT REPORT / NOTESAdjusting OpenMP PageRank : SHORT REPORT / NOTES
Adjusting OpenMP PageRank : SHORT REPORT / NOTES
 
原版制作(swinburne毕业证书)斯威本科技大学毕业证毕业完成信一模一样
原版制作(swinburne毕业证书)斯威本科技大学毕业证毕业完成信一模一样原版制作(swinburne毕业证书)斯威本科技大学毕业证毕业完成信一模一样
原版制作(swinburne毕业证书)斯威本科技大学毕业证毕业完成信一模一样
 
哪里卖(usq毕业证书)南昆士兰大学毕业证研究生文凭证书托福证书原版一模一样
哪里卖(usq毕业证书)南昆士兰大学毕业证研究生文凭证书托福证书原版一模一样哪里卖(usq毕业证书)南昆士兰大学毕业证研究生文凭证书托福证书原版一模一样
哪里卖(usq毕业证书)南昆士兰大学毕业证研究生文凭证书托福证书原版一模一样
 
一比一原版(UofM毕业证)明尼苏达大学毕业证成绩单
一比一原版(UofM毕业证)明尼苏达大学毕业证成绩单一比一原版(UofM毕业证)明尼苏达大学毕业证成绩单
一比一原版(UofM毕业证)明尼苏达大学毕业证成绩单
 
一比一原版(Coventry毕业证书)考文垂大学毕业证如何办理
一比一原版(Coventry毕业证书)考文垂大学毕业证如何办理一比一原版(Coventry毕业证书)考文垂大学毕业证如何办理
一比一原版(Coventry毕业证书)考文垂大学毕业证如何办理
 

[EDBT2021] Assess Queries for Interactive Analysis of Data Cubes

  • 1. EDBT 2021 Assess Queries for Interactive Analysis of Data Cubes Matteo Francia1, Matteo Golfarelli1, Patrick Marcel2, Stefano Rizzi1, Panos Vassiliadis3 1University of Bologna, Italy; 2University of Tours, France; 3University of Ioannina, Greece 24th International Conference on Extending Database Technology (EDBT 2021)
  • 2. EDBT 2021 Motivation Goal: compare business phenomena over multidimensional cubes OLAP provides (low-level) operators to query cubes [1] - Need higher abstractions to express analytic intentions Some operators complement the fundamental ones [2] - E.g., Descriptive [3] or Predictive [4] purposes We introduce the Assess operator - Compare the actual to the expected behavior - Judge the outcome of the comparison Matteo Francia – University of Bologna 2 Introduction [1] Panos Vassiliadis, Patrick Marcel, Stefano Rizzi: Beyond roll-up's and drill-down's: An intentional analytics model to reinvent OLAP. Inf. Syst. 85: 68-91 (2019) [2] Lukasz Golab, Divesh Srivastava: Exploring Data Using Patterns: A Survey and Open Problems. DOLAP@EDBT2021 [3] Sunita Sarawagi: Explaining Differences in Multidimensional Aggregates. VLDB 1999: 42-53 [4] Bee-Chung Chen, Lei Chen, Yi Lin, Raghu Ramakrishnan: Prediction Cubes. VLDB 2005: 982-993
  • 3. EDBT 2021 … in classical OLAP vs IAM Intentional Analytics Model (IAM) [1,2] - Bring OLAP to higher abstraction level - Escape from query answers as sets of tuples Express intentions, not queries - High level syntax - Semantics - Logical/physical optimization with cube [ for predicate ] by levels assess measure [ against <benchmark> ] [ using <function> ] labels 𝜆 Matteo Francia – University of Bologna [1] Panos Vassiliadis, Patrick Marcel, Stefano Rizzi: Beyond roll-up's and drill-down's: An intentional analytics model to reinvent OLAP. Inf. Syst. 85: 68-91 (2019) [2] Antoine Chédin, Matteo Francia, Patrick Marcel, Verónika Peralta, Stefano Rizzi: The Tell-Tale Cube. ADBIS 2020: 204-218 Query the cube Manual comparison - Hard if transformations are needed - Hard with multiple cubes select year, product, sum(quantity) from sales_ft join date_dt on (…) where year = 2019 group by year, product year product sum(quantity) 2019 Milk 125 2019 Apple 100 … … … Introduction
  • 4. EDBT 2021 Computing an Assessment Assess a measure 𝑚 in a cube 𝐶 (target cube) Specify a benchmark cube 𝐵 that - Has a measure 𝑚′ representing the expected behavior of 𝑚 - Is joinable with 𝐶 (𝐵 and 𝐶 might coincide) Optionally apply transformations - E.g., normalization Compare 𝑚 to 𝑚′ cell-wise - E.g., difference and ratio Label cells of 𝐶 based on the comparison - Map the result of the comparison to predefined labels Matteo Francia – University of Bologna 4 𝐶 𝐵 𝑚: quantity = 100 𝑚′: sales = 80 𝐶 C.quantity – B.sales > 0 Assess Operator
  • 5. EDBT 2021 Computing an Assessment 4 benchmark types - Fixed: compare a cube against a fixed value - Sibling: compare two slices from a cube - Past: compare a slice against a historic-based expectation - External: compare two cubes Compare siblings in 𝐶 - E.g., Italian sales are good if higher than French sales with SALES for country = ‘Italy’ by product, country assess quantity against country = ‘France’ using percOfTotal(difference(quantity, benchmark.quantity)) labels {[-inf, -0.2): bad, [-0.2,0.2]: ok, (0.2, inf ]: good} Matteo Francia – University of Bologna 5 Assess Operator SALES Apple Pear Lemon Cola Pizza Beer
  • 6. EDBT 2021 Implementing Assess: 5 Operators Get: given 𝐶, predicate 𝑃, and group-by set 𝐺, returns 𝐶’ Cell-Transform ⊟: cell-at-a-time transformation - Returns 𝐶’ with a new measure value for each cell H-Transform ⊡: holistic transformation - Returns 𝐶’ with a new measure value for each cell given all cells of 𝐶 Matteo Francia – University of Bologna 6 𝐶 𝐶’ 𝐶 𝐶’ 𝐶 𝐶’ Implementation
  • 7. EDBT 2021 Implementing Assess: 5 Operators (Partial) Join ⊠: put together cube and benchmark Pivot ⊞: pivot member’s measure values to new measures Matteo Francia – University of Bologna 7 = ⊠ product ⊞ country, country=‘Italy’ Implementation SALES Lemon quantity=100 COMP Lemon quantity=80 SALES Lemon quantity=100 C’ Lemon quantity=100 quantityFrance=80 = C’ Lemon quantity=100 quantityFrance=80
  • 8. EDBT 2021 SALES Apple Pear Lemon Cola Pizza Beer Sibling Benchmark: Semantics with SALES for country = ‘Italy’ by product, country assess quantity against country = ‘France’ using percOfTotal(difference(quantity, benchmark.quantity)) labels {[-inf, -0.2): bad, [-0.2,0.2]: ok, (0.2, inf ]: good} Get 𝐶 and 𝐵 Partial join on product Transform & label Matteo Francia – University of Bologna 8 Execution Plan Apple Pear Lemon Cola Pizza Beer quantity = 100 quantityFrance = 80 Apple Pear Lemon Cola Pizza Beer quantity=100, quantityFrance=80 diff=20, perc=0.4,label=good
  • 9. EDBT 2021 Can we do better? Intuition: push operators into DBMS to exploit optimization structures From NP to JOP: push cell-wise transformations after join From JOP to POP: transform Join into Pivot Matteo Francia – University of Bologna 9 Naive Plan (NP) Join Optimized Plan (JOP) Pivot Optimized Plan (POP) 1. Get C (SQL) 2. Get B (SQL) 3. Join C and B (memory) 4. Transform (memory) 5. Label (memory) 1. Join C and B (SQL) 2. Transform (memory) 3. Label (memory) 1. Pivot slices in C (SQL) 2. Transform (memory) 3. Label (memory) Execution Plan
  • 10. EDBT 2021 Cube cardinality: SSBi ~= 6𝑀 ⋅ 𝑖 tuples All plans scale linearly Alternative execution plans - JOP faster than NP (when applicable) - POP faster than JOP (when applicable) Past is the heaviest benchmark due to trend estimation Empirical Evaluation Matteo Francia – University of Bologna 10 Evaluation & Conclusion
  • 11. EDBT 2021 Conclusion The assess operator: evaluate cube against benchmark, label the result - Several alternatives for benchmarks, comparison, and labeling schema Alternative execution plans, increase performance by pushing to DBMS Future work - Devise strategies for effectively completing partial assess statements - E.g., if no against is specified, infer the most interesting benchmark - Enhance the expressiveness of the assess - E.g., labeling based on ranges that depend on the coordinates - Investigate operators to develop a cost-based optimization strategy Matteo Francia – University of Bologna 11 Evaluation & Conclusion
  • 12. EDBT 2021 Questions? Matteo Francia – University of Bologna 12 Thank you. Compute query & benchmark Map cells & compare Rule-based cell labeling SALES http://www.cs.uoi.gr/~pvassil/projects/olap_III/