Aster Data Round table

I attended a round table session hosted by TerraData(Aster Data). Here are my learning’s and take aways.

Overview

*Case Study:* Aurora Health Care uses Aster Data to consolidate all their data sources to provide a SQL like interface to their Data set. Behind the scenes, Aster Data can get to any type of data (Big Data(Hadoop), Traditional Data warehouse(Teradata, GP etc..) and any other unstructured or structured data sets.

What is Aster data

Aster looks and acts as a database. Main focus of Aster is on analytics and handles structured semi structured database or MapReduce processes. Assists in Time series analysis, Bask analysis and Full series analysis. It Packages all Analytics into SQL like statements. Aster DOES NOT provide any visualization capabilities.

Aurora health care

* Dave brown is the Director of Business Intelligence from Aurora Health Care in Wisconsin
* A common theme for the hospital is customer centricity
* Analytics team — 30 people are in research. 5 are SAS who are being moved to Aster platform.
* Cheese loving beer drinking population in Wisconsin with wonderful data for sickness.
* 15 days to make data movements from one place to data warehouse. With aster it’s 30 mins.
* BI at ARH is Wired into strategy and planning group and not wired into IT department.
* Tableau was used for 6 years at ARH with great feedback and liking towards that product
* *BI Analytics team controls centrally tableau with data governance and data quality*
* Ownership on data governance was finance and operations since that way they have skin in the game
* 6 yrs ago ARH developed MDM(Master Data Management) around patients
* *Took 3 years to do data governance and needs executive sponsorship.* Highly governed from a data perspective
* *CIO believes that BI is the path to success. CIO level sponsorship for Data Governance, Data Analytics and Data Strategy Roadmap*.

How or What Did ARH do with Aster Data

* Previously ARH always had aggregated data. Now it is granular with Aster Data
* Real time analytics with aster with a vision towards big data analytics. It used to take 15 days to make data movements from one place to data warehouse. With aster it’s 30 mins. Its cheaper better, faster.
* 5 node platform with Aster costs 5th of the mainframe system that they had before.
* 7TB of data in aster system. 
* Aster did the job of educating and evangelizing within the organization.
* ARH uses both MicroStrategy and tableau against aster. (Although based on my question they claimed that they could live with one)
* With Aster Data platform its very Process templates message centric way. Still manage stage and dev environment
* Combing the data with structured and unstructured into one place by aster.

McDonalds

* Top 10 McDonalds around the country has its own data warehouse and they are trying to migrate into global data warehouse
* Per aggregated by market and that data is stored into data warehouse
* Teradata proved that sample is not good enough for McDonalds and hence every level of transaction is being stored now.
* Receives 65 million customers by day worldwide.
* 14 million customers in US which generates huge XML data which is sourced for analytics
* Mobile is more on the brand awarding and DOES NOT make sense for fast food(Drive through mainly).
* Business insights team is by location and also global. More focused on centralizing it globally
* Operations analytics focus is to measure how quick the order was delivered.
* *Customer satisfaction is number one KPI*. Your sales will increases if you beter this KPI. That is how franchise is measured.
* McDonalds is looking into Starbucks way of payment through mobile.

Interesting or Fun Facts

* Big Mac in Japan is 10 dollars and hence needs coupons. It’s in built into Japanese culture that without Coupons they don’t buy things (Source — MCDonald analytics team)
* Loop detectors in the concrete (Drive Through) measures the timing of the service. (Source — MCDonald analytics team)
* The big Hypothesis to detect fraud was: *”Why would anyone go to a drive through to buy coke”*. Was able to detect lot of fraud instances (Source — MCDonald analytics team)
* In Medicaid or Medicare bucket if a patient comes back within 30 days they cannot charge them. Analytics was performed to prevent 1–2% of those readmissions and saved around 60Million (Aurora Health Care BI team)
* Diabetes is curable if caught early (Aurora Health Care)
* Private floors for celebrity/rich patients (sheiks monarchy’s etc..) pays in cash. (Aurora Health Care)
* 20% of geriatrics will come back in 30 days. ARH will not be reimbursed for those patients
* On the same floor as the round table there was another event hosted by Accenture Recruiting. The recruiting event was with an open bar and dinner :-)

My Takeaways

* Everyone is dealing with and trying to solve the same kind of problems. Learn from the industry
* Data Governance and centralized business measurement and monitoring is key to success
* Every tool has its pros and cons. Don’t reinvent the wheel if possible.
* C level sponsorship is needed for data success (In Agile terms they need to be pigs and not chicken)