Presentation is loading. Please wait.

Presentation is loading. Please wait.

Out of the swamp Suggestions to bring your analytics back on track

Similar presentations


Presentation on theme: "Out of the swamp Suggestions to bring your analytics back on track"— Presentation transcript:

1 Out of the swamp Suggestions to bring your analytics back on track
ANALYTICSINSTITUTE problems worth solving Suggestions to bring your analytics back on track MAX COTTICA (Executive M.A.I.I.)

2 Computer Shop Clerk (5 years) IT Development Manager (10 years) SQL DBA/Developer Data Warehouse Developer/Junior Manager (6 years) Data Warehouse Manager About me Global Data Integration Senior Manager (4 years) Head of Data Science And BIG Data Solutions (1 year) 1980s 1990s 2000s 2010s 2017

3 1980 ‘80s ‘90s ‘00s ‘10s Max vs. Analytics in the years: TIER 0 TIER 1
What happened 1980 ‘80s ‘90s ‘00s ‘10s “My goal was to produce reports.” TIER 0 Descriptive Analytics Roll-ups and how, when and where “My goal was to produce ad-hoc reports for a large variety of departments using a centralized media for distribution.” TIER 1 Identify problems and fire alerts Diagnostic Analytics “My goal was to aggregate and integrate data and offer analytics based on different segmentations like time or geography .” TIER 2 Why and forecasting Predictive Analytics “My goal was to use aggregated reports to produce KPI reports and to use predictive models to estimate future growth.” TIER 3 Prescriptive Analytics “My goal is to use predictive analytics in conjunction with scenario based algorithms to produce prescriptive analytics and actionable events.” What will happen if… Next best action

4 Data fit to analytical purpose
Data assurance Data availability

5 We overloaded the lake with raw and meaningless information
That added overhead to the process of discovering, integrating and transforming/aggregating data That also fragmented the tools required to do the job Times for POCs, demonstrations, quick wins, tactical and low hanging fruits went up exponentially due to heavy data wrangling We lost faith in data modelling, data architecture, metadata and data assurance Because of that there has been a huge proliferation of data outside of the lake WHAT happened ?

6 Agile methodologies applied at delivery level and not from the top down
Poor understanding of data assets, no idea where data is Poor or no ODS strategy Lack of product owners, data stewards and data governance Poor master data and reference data management Adoption of a raw data layer with no system refined data Application of security and regulations on top of existing landscape Poor data definitions and integration Lack of data modelling and poor data design and architecture Lack of a metadata and data quality strategy Complex security models Why did it happen ?

7 How do I get back on track ?
Adopt SAFe as a scaled agile methodology, funnel projects don’t sieve them Know your data from source and make sure it is fit for analytical purposes Start a data governance and stewardship program Adoption of data modelling as an enterprise tool, again Adoption of a data assurance strategy Adoption of a raw layer and a refined layer in your logical data warehouse or data lake Adoption of data lineage and metadata capturing Improve data availability Standardize your platforms and development tools Provide better or ODS functionality Consider a dedicated Chief Data Office In a large enterprise consider a dedicated Chief Analytics Office Security and regulations must be part of the data fabric How do I get back on track ?

8 Building your chief data office
Head of Data Regulatory Head of Data Governance and Architecture Head of Data Management and Design Head of Data Security and Integration Head of Analytics and Visualization Risk Data Architects Governance Metadata Security BI & Monitoring Infrastructure Architecture Design EQLT Data Science Stewardship Master Data Big Data MI Data Modellers

9 Building a CHIEF ANALYTICS OFFICE
Machine Learning, Deep Learning, AI, Natural Language Processing, Insights… BI BI on BI Tooling Licensing Monitoring Alerts Functional Data Model Data Science Centralized Code End User Security Scientific Data Model MI Advanced Analytics Standard Reports Predictive, prescriptive, forecast and propensity models

10 Building your execution layer
Data, Master Data and Metadata Management Procurement Integration (EQLT) Consumption Quality Presentation Monitoring Modelling Connection Ingestion Engagement Acquisition Visualization Analytics DATA DEFINITION DOCUMENT Physical Model Functional Model Scientific Model SECURITY MANIFESTO STTM Logical Model Data Architecture and Design Conceptual Model Data Governance and Stewardship

11 BUILDING an analytics or data science team
Feature Catchers SQL Gurus Data Miners Data Wranglers Data Druids Data Scientists Data Analysts Data Engineers Lighthouse Agile leaders Tech Leaders Senior Analysts Senior Scientists Last Guardians Test Engineers QA Engineers CICD

12 Transformed data store
BUILDING YOUR ON PREM HADOOP ecosystem Atomic data store Transformed data store Real-time events CLOUD Star Schemas 1:1 Schema Application transactions EDW Once off EQLT EQLT EQLT ODS 2 MapR-FS VIZ Third party Manual files VIZ MORE VIZ

13 DL NLP ML Building a real time solution
(to disrupt the recruitment industry) CV Mobile ML NLP DL Cover Letter AI BOT Matches Employers Employees LinkedIn Profile WebServices APIs WEB UI Job Specs

14 The data river for analytics: C2A-A2C
NBA NBA Customers NBA NBA NBA NBA NBA NBA NBA Batch NBA Reference Real Time NBA NBA Digital Channel NBA NBA NBA NBA

15 ANALYTICSINSTITUTE problems worth solving


Download ppt "Out of the swamp Suggestions to bring your analytics back on track"

Similar presentations


Ads by Google