#66 How to Structure a Data Analytics Team with June Dershewitz – Director of Analytics

June Dershewitz.jpg

June Dershewitz has spent her career driving analytics strategies for major businesses. She's currently Director of Analytics at Twitch, the world's leading video platform and community for gamers (a subsidiary of Amazon). As an analytics practitioner, she builds and leads teams that focus on marketing analytics, product analytics, business intelligence, and data governance. In her prior life as a consultant, she was a member of the leadership team at Semphonic, a prominent analytics consultancy (now part of Ernst & Young). As a long-standing advocate of the analytics community, she was the co-founder of Web Analytics Wednesdays; she's also a Director Emeritus of the Digital Analytics Association and a current Advisory Board Member at Golden Gate University. She holds a BA in Mathematics from Reed College in Portland, Oregon.

 In this episode, June says she fell into data by accident, she has a bachelor's degree in math. She got an idea that she wanted to move to San Francisco and work with a startup. June was invited for an interview as an analyst, and she loved it. It is only a matter of time for June to meet someone before she starts talking about goats. June grew up on a goat farm; her mother was a farmer. After college, June got a job working for a mathematician in Philadelphia. Later, June decided she wanted to work in the industry. She and her boyfriend moved to San Francisco, and both ended up getting jobs with startups. 

June says in startups; people wear many hats. If you come across a problem, you have the liberty to take ownership and solve it. In smaller companies, it is easier to get a holistic view of the workplace. It is possible at more substantial companies too; June currently works for Twitch, which is owned by Amazon. When June arrived, there were 400 employees, but now has over 1500. One skill June has gotten to use to is stepping up to own things. She still feels empowered to solve problems and to take ownership of those problems. Business intelligence means lots of things to lots of people. Before coming to Twitch, June never claimed to own it outright; however, because she is farther along in her career, she has more experience to draw back on. One risk that we have as data people is potentially getting pigeonholed into a thing that we are good at and getting stuck there. June knows a ton about Adobe Analytics; she could have continued to exist in that realm and be just fine. However, she thrives on solving a multitude of problems and being challenged in the workplace. Later, June discussed the hub and spoke organization model, the data quality journey at Twitch, and getting involved in the data science community. 

Enjoy the show!

We speak about:

  • [01:40] How June started in the data space

  • [08:20] Solving problems in startups 

  • [09:45] Getting a holistic view in the workplace  

  • [11:20] Feeling unsure about owning a piece of work 

  • [15:30] Business intelligence skillsets for data scientists   

  • [19:35] Clear understanding of data roles in the workplace 

  • [20:55] An overview of June’s teams’ structures 

  • [27:10] Managing career transitions with the hub and spoke model  

  • [29:25] Assigning each person a technical buddy  

  • [32:10] The data quality journey  

  • [41:40] Evolution of data quality at Twitch  

  • [48:00] Becoming involved in the data science community  

  • [53:10] Other ways June stays involved in her communities 

  • [55:20] Advice for breaking into the data science field 

Resources:

June’s LinkedIn: https://www.linkedin.com/in/jdersh

Non-Invasive Data Governance: The Path of Least Resistance and Greatest Success

Data Governance: How to Design, Deploy and Sustain an Effective Data Governance Program

Quotes:

  • “The thing about being a data person at that time was we just had to figure it out.”

  • “I was the vice president of everything that needed to get done.”

  • “At Twitch, we don’t have a clear definition of what a data engineer means.”

  • “We chose to move to an organization model that is hub and spoke.”

  • “Data governance can mean lots of things to lots of people.”

Thank you to our sponsors:

UNSW Master of Data Science Online: studyonline.unsw.edu.au

Datasource Services: datasourceservices.com.au or email Will Howard on will@datasourceservices.com.au

Fyrebox - Make Your Own Quiz!

June Dershewitz is based in San Francisco, California.

And as always, we appreciate your Reviews, Follows, Likes, Shares and Ratings. Thank you so much for listening. Enjoy the show!

#64 Intersections of Analytics, AI, Linguistics and Culture with Prashant Natarajan – Principal, AI & Analytics

Prashant Natarajan.jpeg

Prashant Natarajan has 18+ years’ experience in building EMRs, ERP, big data platforms, actionable analytics, and machine/deep learning applications. Before joining Deloitte, he served in hands-on global consulting and product leadership roles at H2O.ai, Oracle, McKesson Payer Solutions, Healthways, and Siemens. Prashant is Co-Faculty Instructor of Data Science and AI at Stanford University School of Medicine, Palo Alto, CA, USA. He volunteers as an industry expert and guest lecturer at leading Australian universities. Prashant serves as an industry advisor at the CIAPM computer vision project in University of California San Francisco, Council for Affordable Health Coverage, and Pistoia Alliance Center for Excellence in Artificial Intelligence.  

In this episode, Prashant describes how essential human interaction is for success. In a technology-heavy space, human interaction and linguistics were not very common. Instead of complaining about it, Prashant went and got his masters to focus on English in the technology space. To have success, we need a clear understanding of culture. Culture is language, and language at its core is mathematics. How do we interact with people to figure out what their strengths are? Prashant considers himself the luckiest person on earth to have the experiences he has had in his career. 

Then, Prashant discusses how to identify business problems and integrate it with data science. Data analytics is as old as when humans started interacting with each other. By nature, all human beings are data analytics, consuming creatures. Historically, the application of computing data and analysis have been humans trying to define a problem, find data to solve the problem, and then writing algorithms that will create insights. Prashant is a massive admirer of the human mind and the human brain, which is far superior to any artificial machine. When it comes to AI, computing can do things that the human mind cannot. Business leaders must look at data science as a way to help them define business problems, rather than purely using deduction. Later, Prashant advises companies moving into data-driven products, explains horizontal capabilities, and the use of machine learning in healthcare. 

Enjoy the show!

We speak about:

  • [01:25] How Prashant started in the data space 

  • [03:45] Studying communications and linguistics  

  • [08:45] Mentoring young professionals  

  • [11:45] Work with people who are smarter than you  

  • [15:00] Merging business problems with data science  

  • [19:45] The value business leaders see in data  

  • [25:00] Advice for companies who are moving into data-driven products

  • [29:45] What excites Prashant about the future of data 

  • [34:05] Horizontal capabilities 

  • [37:20] The use of machine learning in healthcare  

  • [44:20] Improving product development 

  • [48:40] Prashant’s proudest moment

  • [50:15] The manufacturing industry 

  • [52:20] We learn more from our failures than our successes 

Resources:

Prashant’s LinkedIn: https://www.linkedin.com/in/natarpr/

Demystifying Big Data and Machine Learning for Healthcare (Himss Book)

Quotes:

  • “Human interaction is the most key determiner of success or not.”

  • “Today, we have the technology that has caught up with the human need.”

  • “Data science is increasingly a horizontal capability that will impact all of us.”

  • “I celebrate relationships because they allow me to learn.”

Now you can support Data Futurology on Patreon!

https://www.patreon.com/datafuturology

Thank you to our sponsors:

UNSW Master of Data Science Online: studyonline.unsw.edu.au

Datasource Services: datasourceservices.com.au or email Will Howard on will@datasourceservices.com.au

Fyrebox - Make Your Own Quiz!

Prashant Natarajan is based in Melbourne, Victoria, Australia.

And as always, we appreciate your Reviews, Follows, Likes, Shares and Ratings. Thank you so much for listening. Enjoy the show!

#62 Full Stack Data Science with Gregory Hill – Global Head of Analytics

Gregory Hill.jpg

Dr. Gregory Hill leads the Analytics function at Brightstar's Global Services division, developing and delivering their data & analytics strategy, innovation programs, and product development initiatives. He works across their lines of business, including supply chain optimization, product portfolio management, financial services, buy-back and trade-in, leasing, and omnichannel solutions. He also manages Brightstar's analytics team in support of their key global accounts with pre-sales, solution design, and service delivery. His expertise is in the application of advanced analytics techniques (including machine learning, predictive modelling, mathematical optimization, econometrics, and operations research) to commercial problems. These applications span forecasting, pricing, fraud, market segmentation, customer satisfaction, and propensity modelling.

In this episode, Gregory explains how he started in the data space. He was aware of all the theoretical work being done around data but did not know how it worked in an industry aspect. The real challenge of putting mathematical models to practice lies in the organizational and people elements of it. Computer science and electrical engineering do not teach you how to overcome organizational challenges and individual motivations and incentives. Going back to get his Ph.D., Greg wanted to do something requiring qualitative research. So he targeted informational systems and economics. His fieldwork leads him to interview executives of larger banks, publicly listed companies, and government agencies. He came up with an economic framework that improved customer data quality. 

Some problems Greg started looking into while working at Telstra were fixed by using the four P's of marketing. He had an opportunity to learn something new that they did not teach in engineering school. In business, the four P’s are a useful lens to think about commercial problems like product lifecycle management and portfolio optimization for business. They looked at questions around what products will work well in what channels. Previously, this type of merchandising decision making was done by gut feel. Having a data-driven approach was a different way of thinking for the company and the teams. Greg would not recommend someone to gain a Ph.D. to become a data scientist. You can acquire the skills you need outside of academia. Academia will not give you the skills to become successful, a Ph.D. may hinder requiring all the skills to become a data scientist. Later, Greg discusses his appreciation for managing data scientists, being involved in the local data community, and the challenges of working globally. 

Enjoy the show!

We speak about:

  • [02:00] How Greg started in the data space

  • [11:10] Leaving academics and getting involved in the industry  

  • [13:20] Greg’s work background

  • [18:25] The four P’s of marketing

  • [20:40] Transiting from gut instinct to a data-driven approach

  • [27:55] Thinking through cause and effect 

  • [30:45] What Greg’s team looks like

  • [39:00] Lessons learned from managing data scientists  

  • [42:25] Active in local data science meetups + guest speaking  

  • [44:25] Working globally + peeling back opportunities to use data science techniques

Resources:

Greg’s LinkedIn: https://www.linkedin.com/in/gregoryhill/?originalSubdomain=au

Brightstar: https://www.brightstar.com

Quotes:

  • “My thesis was not a project; it was a lifestyle.”

  • “I didn’t want to be an academic, I wanted to get back into the industry.”

  • “It was a combination of arrogance and laziness.”

  • “At the end of the day, it boils down to if I change X, will Y change?”

Now you can support Data Futurology on Patreon!

https://www.patreon.com/datafuturology

Thank you to our sponsors:

UNSW Master of Data Science Online: studyonline.unsw.edu.au

Datasource Services: datasourceservices.com.au or email Will Howard on will@datasourceservices.com.au

Fyrebox - Make Your Own Quiz!

Gregory Hill is based in Melbourne, Australia.

And as always, we appreciate your Reviews, Follows, Likes, Shares and Ratings. Thank you so much for listening. Enjoy the show!

#59 Creating the Link Between Business and Data with Tony Gruebner - GM Analytics, Insights and Modelling

Tony Gruebner.jpg

Tony Gruebner is the GM Analytics of Insights and Modelling and the Exec Sponsor of Personalisation at Sportsbet. He established a department of 40+ skilled analysts and data scientists tasked with creating innovative data products focused at improving the experience for their customers and supporting the business by providing relevant and timely information and insights that steer decision making across all levels of the business. He has served on the Executive Leadership Team from 2016.

In this episode, Tony explains how he started in data and what led him to get his job at Sportsbet. Tony got a call from a recruiter asking if he wanted to do work with analytics, in a company that does sports and is heavily digital. All of those factors checked the box for Tony, and he took the entry-level analyst role. Over time, the need for analytics has grown, so he has been able to develop some analytics teams. 

One thing Tony does with Sportsbet is getting his data scientists to understand the link with the business, and the company to understand the connection with data science. The main strength of a data scientist is creating models; however, you need someone from the business side to tell them what kind of problems they need to be solved. Then, the data scientists can work on how to solve the issues numerically. Also, the business can consider things they currently cannot do, but data scientists can enable them to do these things. For Sportsbet, everything they do is to improve the customer experience.

Then, Tony explains how his team communicates how data science works. They actually painted out what the modelling cycle looked like and presented it to a lot of people. The feedback he got back is people understood the process and where the difficulties lie in the process. Communicating this always is challenging, especially when embarking on a project that may take six to twelve months to accomplish. Tony suggests breaking up the project into little chunks to avoid miscommunications. In one instance, after six months of work, the company explained the data was not solving the problem they wanted to be addressed. If Tony broke up the project into smaller pieces, this could have been avoided.

Later, Tony explains how Sportsbet is trying to scale globally and all the nuisances that come with the territory. For instance, they need to figure out how to acquire global talent and to overcome uncomplimentary time zones. Also, his team is working on how to utilize artificial intelligence to solve problems. Some of the models work on improving the customer experience directly. Just like how Netflix recommends movies, they are working on recommending a specific horse race to the customer; however, when the race is done, it needs to disappear from the site. Whereas, when Netflix recommends a movie to a customer, it can stay there for potentially years. Sportsbet is working on artificial intelligence to improve these models for their consumers. 

Enjoy the show!


We speak about:

  • [01:20] How Tony got started in data

  • [08:20] Tony’s skills come from the commercial side

  • [11:10] Linking data science and the business

  • [14:30] Communicating how data science works

  • [17:00] Steps to getting others to understand data science

  • [20:40] Getting the best talent for your team

  • [24:00] Structuring teams and the department

  • [28:10] Transiting from analytical roles to commercial roles 

  • [35:30] Working on global expansion

  • [38:10] Solving with artificial intelligence

  • [42:30] Passionate about using numbers to reach an outcome

  • [44:00] Modelling failures with Sportsbet 

  • [47:50] Imposter syndrome in data science  

  • [50:05] Data science is rapidly changing and exciting

Resources:

Tony’s LinkedIn: https://www.linkedin.com/in/gruebz/

Sportsbet: https://www.sportsbet.com.au

Tony’s Twitter: https://twitter.com/gruebz?lang=en

Quotes:

  • “There is no one path that always works.”

  • “There are literally thousands of things data scientists couldn’t potentially tackle in any business.”

  • “If you’re not making mistakes, then you aren’t pushing the envelope hard enough.”

  • “Not having imposter syndrome is a sign of lack of knowledge.”

Now you can support Data Futurology on Patreon!  

https://www.patreon.com/datafuturology 

Thank you to our sponsors: 

UNSW Master of Data Science Online: studyonline.unsw.edu.au 

Datasource Services: datasourceservices.com.au or email Will Howard on will@datasourceservices.com.au 

Fyrebox - Make Your Own Quiz!

Tony Gruebner is based in Melbourne, Victoria, Australia.

And as always, we appreciate your Reviews, Follows, Likes, Shares and Ratings. Thank you so much for listening. Enjoy the show!

#57 How To Go From Academic to Data Science Leader with Yuval Marom - Analytics and Data Science Professional

Yuval Marom.jpg

Yuval is an Analytics and Data Science professional with extensive commercial and academic experience. His interests and goals are to be working on interesting and practical problems where there is a need to discover and act on meaningful patterns in data, through advanced analytics and data science. I'm the founder and co-organiser of two meetups: Data Science Melbourne and MelbURN, a user group for Melbourne-based users of the R statistical and data mining programming language. 

In this episode, Yuval tells us about how both of his parents are statisticians and inspired him to fall in love with data science. Growing up, he used Pascal to build spaceship games, and it motivated his passion for programming. Eventually, Yuval went for his Ph.D. and focused on applying how animals learn and behave to robotics. Simulated and physical experiments were pretty basic because robotics were not as advanced as they are today. Later, Yuval realized academia was not necessarily his calling, he was more interested in applying solutions to interesting problems. However, in recent years, research innovation and solving problems are becoming much more intertwined. 

Then, Yuval tells us why it essential to embrace simplicity and praises the advantages he reaped by working for a small business. In retrospect, Yuval realized he developed a higher level perspective of business level thinking by working for such a small team; however, he actually turned down a promotion because he was not yet confident in his technical skills. Get as close to the source of data as possible to get a real appreciation for it. Without appreciation, a person will not be successful in a data science managerial position. Doing technical work, Yuval was able to stay connected to the community and hear varying perspectives on how businesses operate and the different flavors of management. Later, Yuval explains his struggles with becoming a manager, the benefits of building connections in the workplace, and the importance of allocating time for professional development. 

 Enjoy the show!

We speak about:

  • [01:40] How Yuval fell in love with data science

  • [05:45] Social learning in biology

  • [08:05] Lessons learned from completing a Ph.D.

  • [13:10] Research innovation vs. solving problems

  • [15:40] Embrace simplicity 

  • [18:00] Small business advantages 

  • [21:45] Skills to develop before management 

  • [26:00] Results oriented work

  • [30:45] Different flavors of management

  • [32:50] Connection to community 

  • [40:20] Learning to interact with stakeholders + managerial skills 

  • [44:00] Benefits of building connections + education 

  • [48:00] Assume people are at work with good intentions 

  • [52:00] Allocate time for professional development 

  • [59:30] Focus on retention

Resources:

Data Science Melbourne 

MelbURN

Yuval’s LinkedIn

University of New South Wales

Tweetable Quotes:

  • “Embrace simplicity or go to academia.”

  • “You have to know the business process before you can make sense of the data.”

  • “Some people strive from innovation, whereas others find satisfaction from solving simple problems and making a difference in an organization.”

  • “If you have the right relationships, you can make anything work.”

  • “It’s important to try different experiences while you’re young, but facing the challenges where you are is equally important.”


Now you can support Data Futurology on Patreon!  

https://www.patreon.com/datafuturology 

Thank you to our sponsors: 

UNSW Master of Data Science Online: studyonline.unsw.edu.au 

Datasource Services: datasourceservices.com.au or email Will Howard on will@datasourceservices.com.au 

Fyrebox - Make Your Own Quiz!

Yuval Marom is based in Melbourne, Australia.

And as always, we appreciate your Reviews, Follows, Likes, Shares and Ratings. Thank you so much for listening. Enjoy the show!



#50 Prakash Baskar – Founder and President

Prakash Baskar.jpg

Prakash Baskar is the Founder and President of Khyanafi. He helps data leaders to rapidly transition and accelerate the success of data, analytics, and digital initiatives. Previously, Prakash was the Chief Data Officer at Santander Consumer USA where he led enterprise data governance, risk infrastructure & information (risk data aggregation), data quality, business data strategy & solutions, and business & reporting analysis functions.

In this episode, Prakash tells us how he started in the data space at his university. His role was to determine how students were performing. If they are not performing well, he needed to identify why. The graduation rates were low at the school, so Prakash was tasked with finding out what was the problem. Then, Prakash discusses starting a new job and having little direction about what to do. With everchanging technology, the description of your job will always be changing too. As a person going into any role, understand that you do not have to ask permission all the time. Have a clear idea of what you can do and what you cannot do, then do what you feel is right for the organization. Look for where the opportunities for expansion are and find a way to get results.

If you ask ten people what the role of a Chief Data Officer is, you will get ten different answers. Whatever the CDO does will ultimately be to enable others to receive real benefits out of the data. Just because something is not broken, does not mean it cannot be improved. There are many different routes a person can take to become a CDO; however, you need someone with knowledge in multiple aspects of business, technology, and people management. A CDO needs to create value for the organization; learn the company you are supporting to anticipate the problems they may run into.

Later, Prakash explains how in business, any change is hard. How you embrace the change after it is made is what will differentiate yourself from others. If the change is too complicated, people will shut off. Start off by telling the client what the change will do for them rather than the steps it will take to get there. Some other tips when presenting a significant change is to be realistic with what it will take and make sure not to overpromise. It is imperative to select things that you can quickly do with minimal engagement from their people. Plus, make sure you have updates for the company each month, so they understand what is being revealed from the data. Finally, Prakash discusses how essential it is to move around the organization in order to understand different departments and he reveals the inspiration behind his latest business venture.

Enjoy the show!

We speak about:

  • [01:30] How Prakash started in the data space

  • [03:55] The transition into consulting  

  • [07:50] Deciding what problems to tackle first on the job

  • [10:10] The role of Chief Data Officer  

  • [17:20] Creating value for the organization  

  • [22:10] Businesses getting the maximum benefit from analytical work

  • [31:25] How to determine what to work on first with a company

  • [39:00] Data science conferences are full of CDOs

  • [45:45] Actively moving around the organization

  • [50:20] The inspiration behind Khyanafi

  • [53:50] What do you think makes a great leader in the data space?

  • [55:40] Advice for data scientists   

Resources:

Prakash’s LinkedIn: https://www.linkedin.com/in/prakashbaskar/

Khyanafi: http://www.khyanafi.com

Khyanafi’s LinkedIn: https://www.linkedin.com/company/khyanafi/about/

Set up a discovery call with Prakash: https://calendly.com/prakashbaskar/discoverycall

Quotes:

  • “Your job description is only as good as the time it was written.”

  • “I am not a big believer in five or ten-year plans; I make plans six months at a time.”

  • “CDOs have to know the business! Try and learn how the company operates; you will gain more respect when you take the effort to understand the business.”

Now you can support Data Futurology on Patreon!  

https://www.patreon.com/datafuturology 


Thank you to our sponsors: 

UNSW Master of Data Science Online: studyonline.unsw.edu.au 

Datasource Services: datasourceservices.com.au or email Will Howard on will@datasourceservices.com.au 

Fyrebox - Make Your Own Quiz!

Prakash Baskar is based in Columbus, Ohio.


And as always, we appreciate your Reviews, Follows, Likes, Shares and Ratings. Thank you so much for listening. Enjoy the show!