2 12.07.2017 University of Massachusetts Amherst Boston - - PowerPoint PPT Presentation

▶
2
SMART_READER_LITE
LIVE PREVIEW

2 12.07.2017 University of Massachusetts Amherst Boston - - PowerPoint PPT Presentation

Nov 23, 2023 229 likes •453 views

University of Massachusetts University of Massachusetts Amherst Boston Dartmouth Lowell Worcester UMassOnline Amherst Boston Dartmouth Lowell Worcester UMassOnline Welcome! 2 12.07.2017 University of Massachusetts

slide-1
SLIDE 1

University of Massachusetts

Amherst Boston Dartmouth Lowell Worcester UMassOnline

University of Massachusetts

Amherst Boston Dartmouth Lowell Worcester UMassOnline

2

12.07.2017

Welcome!

slide-2
SLIDE 2

University of Massachusetts

Amherst Boston Dartmouth Lowell Worcester UMassOnline

Today’s Agenda

Topic Presenter Length Start End Welcome and Agenda Shahr Panahi :10 9:30 9:40 Introduction/Keynote John Letchford :25 9:40 10:05 Summit Status and Roadmap Shahr Panahi :40 10:05 10:45 A&F Dashboard Discussion Lisa Calise :30 10:45 11:15 Break :15 11:15 11:30 BPR Update BPR Leads :25 11:30 11:55 Tableau for Summit Bill Manteiga :15 11:55 12:10 HelioCampus Introduction Lori Dembowitz :20 12:10 12:30 Lunch (Provided Downstairs) 1:00 12:30 1:30 Breakout Sessions :HR Carol Dugard, HR Attendees 1:30 1:30 3:00 :Finance John Munroe, Finance Attendees 1:30 1:30 3:00 :Student Jeff Glatstein, Student Attendees 1:30 1:30 3:00 Report back from Breakouts All :30 3:00 3:30 Closing Remarks/Evaluation forms Shahr Panahi :10 3:30 3:40 Middlesex Essex Berkshire

slide-3
SLIDE 3

University of Massachusetts

Amherst Boston Dartmouth Lowell Worcester UMassOnline

University of Massachusetts

Amherst Boston Dartmouth Lowell Worcester UMassOnline

Roadmap

DECEMBER 2017

SUMMIT Summit

By Anju Sherpa - Own work, CC BY-SA 3.0, https://commons.wikimedia.org/w/index.php?curid=33073777

slide-4
SLIDE 4

University of Massachusetts

Amherst Boston Dartmouth Lowell Worcester UMassOnline

Topics

– What is Summit? – Where are we today:

  • Architecture
  • Usage

– Where are we going

  • UMass Community of Data

Practitioners

  • Future Architecture (Draft)
  • Roadmap
slide-5
SLIDE 5

University of Massachusetts

Amherst Boston Dartmouth Lowell Worcester UMassOnline

What is Summit?

  • UMass’ Business Intelligence (BI) and Analytics program.
  • The main objective of Summit is to facilitate and promote data-centric decision

making by:

  • Deploying solutions that transform raw data into actionable information;
  • Providing access to that information to decision makers;
  • Championing data governance across the university;
  • Supporting BI / Analytics community’s data, information, and technology needs.

– UMass Enterprise Data Architecture

  • Enterprise Data Warehouse
  • Enterprise Reporting and Analysis tools (OBIEE, Tableau, etc.)
  • Data Integration for Analysis – Tools that transform and move data
  • Data Access for all analysts
  • Future technologies such as “data lake”, access brokers, Analytics marts
slide-6
SLIDE 6

University of Massachusetts

Amherst Boston Dartmouth Lowell Worcester UMassOnline

University of Massachusetts

Amherst Boston Dartmouth Lowell Worcester UMassOnline

SUMMIT: Today’s Architecture

Summit Data Marts EDW

PS – SA

BDL

IR Census Mart

PS HCM PS Fin

Buyways, Sunapsis, Equifax, IDM, Other Sources

3rd Party Integration

Access Layer: OBIEE

External Data

Campus Marts

Metadata Layer

Campus Marts Tableau Server Tableau Tableau Tableau Tableau Tableau Tableau Servers PS – SA

AMHERST

PS – SA

MEDICAL

slide-7
SLIDE 7

University of Massachusetts

Amherst Boston Dartmouth Lowell Worcester UMassOnline

Where are we: Summit Usage and Statistics

  • Increasing amount of data

being queried (800 M rows per quarter)

  • 17 Data Marts, 40 ++

dashboards

  • Thousands of unique

users per month

slide-8
SLIDE 8

University of Massachusetts

Amherst Boston Dartmouth Lowell Worcester UMassOnline

How do we as a university system maximize potential of data analytics in the most efficient and productive way?

– People, Organization, Processes

  • Position UMass resources on campus and centrally for maximum efficiency and

productivity

  • Invest in analysts on campus to answer business questions using data
  • Centralize where it can lead to efficiencies

– Technology and Architecture

  • Modernize UMass enterprise data architecture
  • Take advantage of new technologies, cloud hosting, best practices in analytics

– BI / Analytics Content

  • Build / buy analytics content to support UMass strategic direction both on campus and

as a system

  • Care for campus as well as system data and analytics needs
slide-9
SLIDE 9

University of Massachusetts

Amherst Boston Dartmouth Lowell Worcester UMassOnline

UMCDP

UMass Community of Data Practitioners

UMass Community of Data Practitioners

Introducing UMass Community of Data Practitioners

UMASS COMMUNITY OF DATA PRACTITIONERS:

  • Collection of BI / Analytics professionals across

UMass

  • Optimize collaboration across campuses and UITS

NOW:

  • Share knowledge;
  • Collaborate online
  • Organize events

FUTURE:

  • Project and research opportunities for faculty and

students

  • Budgeted through grants where possible
  • Cooperate with higher ed. Institutions across the

commonwealth and beyond

  • Invest in and provide access to new technologies

around data

  • Support open source innovation
  • Involved in data governance and strategy for the

enterprise

  • Kaggle like analytics contests and scholarships on

real life UMass business questions

AMH BOS CEN/ UITS DAR LOW MED

slide-10
SLIDE 10

University of Massachusetts

Amherst Boston Dartmouth Lowell Worcester UMassOnline

Future Summit Architecture Guiding Principles

  • Facilitate ease of access to all available data for mining and analysis
  • Strive to make transforming data into actionable information ,including advanced analytics

simple and repeatable

  • Accommodate purchased analytics solutions / hosts
  • Support governance, standardization, and access security
  • Care for agility and self service in developing BI / Analytics content
  • Take advantage of latest technology and cloud hosting
  • Pay special attention to user experience, make mobile available
  • Ensure proper buy-in and support from UMass BI / Analytics community
  • Capitalize on the community and engage faculty and student practitioners
  • Prioritize based on utility and benefit to the entire enterprise (rather than single campus)
  • Maximize use of existing investments and minimize ‘redoing’ work that has already been

done

slide-11
SLIDE 11

University of Massachusetts

Amherst Boston Dartmouth Lowell Worcester UMassOnline

University of Massachusetts

Amherst Boston Dartmouth Lowell Worcester UMassOnline

Analytics Engine Artificial Intelligence

Access Layer: Reporting and Analytics tools Optimized and integrated for best user experience

Enterprise Data Warehouse

Highly Cleansed Transformed Data

Future state enables big data, analytics, self service, mobile, and data exploration Data Lake

Massive Repository of Raw Data In All Formats Sources: All ERP, Cloud, On Premise, and External Data Sources

Metadata

Entire Stack Hosted on Cloud?

Summit: Future BI/Analytics Architecture

Access Broker

Source Layer Repository Layer Marts Layer Access Layer

Data Integration

Other Web & Mobile Applications

Specific Data Marts (Collection of Cleansed Data for Specific Subjects) Content Vendors (e.g. HelioCampus)

slide-12
SLIDE 12

University of Massachusetts

Amherst Boston Dartmouth Lowell Worcester UMassOnline

What is a Data Lake

  • Able to store vast quantities of data in raw format
  • It stores data of different types:

– Database tables – Log files (i.e. web log information) – Binary files (i.e. pictures, voice, etc.) – Other

  • It can be Hadoop based, RDBMS based or both
  • Used for Advanced analytics as well as quick access to data (don’t have to

wait for data to get to EDW before using)

  • Best practice* is to implement alongside enterprise data warehouse

* According to TDWI, Gartner, Cloudera

slide-13
SLIDE 13

University of Massachusetts

Amherst Boston Dartmouth Lowell Worcester UMassOnline

Architecture Roadmap

Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2

2017 2018 2019 2020 2021

Oracle 12.2 / HW Upgrade OBIEE 12c / HW Upgrade

Tableau

Metadata Integration HelioCampus Deployment (DL) Enterprise Server (PaaS) Cloud Host Oracle Cloud? More campus /Central deployments (ABWC) ?? Placeholder Planned Milestone Decision

Helio Campus EDW OBIEE

Expand Central Server (PaaS?) Vendor Acquisition & Deployment

Data

Lake / Access / Auto-Discovery

Discovery Tool

Advanced Analytics

First Predictive Model Advanced Analytics Development SQL Server POC

slide-14
SLIDE 14

University of Massachusetts

Amherst Boston Dartmouth Lowell Worcester UMassOnline

Content Roadmap

  • Some key upcoming content:

– A&F Executive Dashboard – A&F Executive Dashboard: Campus Detail – System IR Data Mart – Deans’ dashboard – HelioCampus Visualizations – Student Success Appliance – Predictive models – …

slide-15
SLIDE 15

University of Massachusetts

Amherst Boston Dartmouth Lowell Worcester UMassOnline

Appendix

slide-16
SLIDE 16

University of Massachusetts

Amherst Boston Dartmouth Lowell Worcester UMassOnline

Basics: Optimizing Data Analysis and Delivery

Traditional

Data Discovery and Acquisition

Time and Effort

Data Cleansing and Transformation Analysis Information Delivery

Majority of the time and effort by analysts is spent on accessing and preparing

  • data. This reduces time

available for actual analysis!

slide-17
SLIDE 17

University of Massachusetts

Amherst Boston Dartmouth Lowell Worcester UMassOnline

Basics: Optimizing Data Analysis and Delivery

Traditional

Data Discovery and Acquisition

Time and Effort

Data Cleansing and Transformation Analysis Information Delivery

Optimized

Tools to access data from disparate systems, raw data repositories (like data lakes), automated discovery tools will help access and discovery

  • Simplify governed data access
  • Invest in metadata
  • Automate discovery and visualization
  • Make ALL data reachable for analysis

(Data Lake + Data Warehouse + …)

slide-18
SLIDE 18

University of Massachusetts

Amherst Boston Dartmouth Lowell Worcester UMassOnline

Basics: Optimizing Data Analysis and Delivery

Traditional

Data Discovery and Acquisition

Time and Effort

Data Cleansing and Transformation Analysis Information Delivery

Optimized

  • Use EDW and marts where possible
  • Automate data wrangling
  • Standardize data definitions

The EDW already has huge amounts of cleansed data; subject specific marts can help in automating data transformation and cleansing

  • Simplify governed data access
  • Invest in metadata
  • Automate discovery and visualization
  • Make ALL data reachable for analysis

(Data Lake + Data Warehouse + …)

slide-19
SLIDE 19

University of Massachusetts

Amherst Boston Dartmouth Lowell Worcester UMassOnline

Basics: Optimizing Data Analysis and Delivery

Traditional

Data Discovery and Acquisition

Time and Effort

Data Cleansing and Transformation Analysis Information Delivery

Optimized

  • Use EDW and marts where possible
  • Automate data wrangling
  • Standardize data definitions

We will have more time to do analytics, and more time to spend with business

  • wners, defining questions

and refining results. Where feasible, we can ‘buy’ rather than ‘build’. Campus (and central) analysts can spend more time on the actual analysis.

  • Simplify governed data access
  • Invest in metadata
  • Automate discovery and visualization
  • Make ALL data reachable for analysis

(Data Lake + Data Warehouse + …)

  • Enable best of breed BI tools
  • Invest in data visualization
  • Open source and other advanced

analytics

  • Analytics marts and servers
  • Facilitate self service on campus
slide-20
SLIDE 20

University of Massachusetts

Amherst Boston Dartmouth Lowell Worcester UMassOnline

Basics: Optimizing Data Analysis and Delivery

Traditional

Data Discovery and Acquisition

Time and Effort

Data Cleansing and Transformation Analysis Information Delivery

Optimized

  • Use EDW and marts where possible
  • Automate data wrangling
  • Standardize data definitions
  • Simplify and automate deployment of

interactive content and advanced analytics results

  • Streamline user experience, add mobile

By automating the delivery

  • f analytics results:

interactive dashboards, visualizations, and/or model scores, we are able further

  • ptimize the process. Pay

attention to overall user experience.

  • Simplify governed data access
  • Invest in metadata
  • Automate discovery and visualization
  • Make ALL data reachable for analysis

(Data Lake + Data Warehouse + …)

  • Enable best of breed BI tools
  • Invest in data visualization
  • Open source and other advanced

analytics

  • Analytics marts and servers
  • Facilitate self service on campus
slide-21
SLIDE 21

University of Massachusetts

Amherst Boston Dartmouth Lowell Worcester UMassOnline

University of Massachusetts

Amherst Boston Dartmouth Lowell Worcester UMassOnline

Analytics Marts Analytics Marts

Enterprise Data Warehouse

OBIEE Tableau

Analytics AI Engine

Future state enables big data, analytics, self service, mobile, and data exploration Data Lake

AWS, Azure, Cloudera, Oracle or …?

Analytics Marts Subject Marts Advanced Analytics LMS Other ERP

Other Cloud

SalesForce

Metadata

Entire Stack Hosted on Cloud?

Future BI/Analytics Architecture

Other BI Tools

Access Broker

Source Layer Repository Layer Marts Layer Access Layer Mobile

Data Integration

Other Web & Mobile Applications

Web Integration for Seamless User Experience

Metadata Content Vendor(s) Content Vendor(s) e.g. HelioCampus