Data Science is No Longer Just for Data Scientists
In this podcast, Imad Birouty, Director of Technical Product Marketing with Teradata, discusses the democratization of data science to enable users across the organization to find the answers on which to base their decisions. He explains how Teradata is making is easier to give users access to machine learning and advanced analytics. The interview is conducted by Phil Bowermaster, an independent consultant and analyst who writes and speaks about emerging technologies and the future.
For years we have heard about democratizing data, but now we’re hearing about the idea of democratizing data science. What does that mean?
Imad Birouty: It is an interesting term. The idea of democratizing data science is to make the data science process available to more people throughout the organization – the knowledge workers, the business analysts. In doing so, it’s important to separate the data scientist from data science. The data scientist is the person – it’s a job role. Data scientists have lots of different skills and tools that they use.
The data science process is more of a discovery process – finding interesting patterns in data. You can’t limit that to just a single job role. Every person does that in their daily jobs. Everyone has to make decisions. Everyone needs data. So the idea of democratizing a data science process is a good thing. If you look at an entire organization from the CEO on down to the very lowest level of the organization, everyone needs to make decisions. Some decisions are more important than others, but people need informed data to make decisions. Intelligent people will follow intelligent processes to substantiate their decisions. The process of democratizing data science is to make the tools and the capabilities available to a broader range of users.
Years ago I was involved in marketing an analytics product that no longer exists, but one of the taglines we came up with at the time – we thought it was pretty clever – was: Everybody gets to be Einstein. Is that what’s happening here? Is everybody getting to be a data scientist?
Imad Birouty: If done right, everyone in the organization can benefit, but really the largest benefactors of this would be the business analysts, knowledge workers or analysts – every organization uses slightly different titles. For example, maybe someone in inventory has to decide how much inventory to stock and determine if stocking lower levels would bring the risk of running out of stock. A human being – a knowledge worker – has to make that decision. It is knowledge workers – whether you call them business analysts, knowledge workers, or analysts – who have to make these kinds of decisions. They understand their business, they understand their company, and they are making important decisions that they need to back up with information.
Those types of knowledge workers are going to be the ones that benefit the most from having advanced analytics,
machine learning and different algorithms supporting and augmenting their decision-making process. They’re the real benefactors.
So you take the toolkit of the data scientists and put it into the hands of the business analysts/knowledge workers. That sounds great in theory, but what is Teradata doing to make that happen?
Imad Birouty: We’re making advanced analytics and machine learning much more accessible. A few years ago, Teradata bought a company called Aster. One of the things that Aster was able to do was to take the machine- learning functions and wrap them in SQL. More people in organizations know how to use SQL versus those who know how to code in third-generation procedural languages. SQL is more the language of analysts.
That was a first step. Taking advanced analytics and machine learning and wrapping them in SQL was a step in the right direction. Now if you want to expand it even more, the next step in making this easier to use by a broader majority is adding a nice point-and-click interface. To accomplish that, we just announced a product called Vantage Analyst.
Vantage Analyst takes a host of machine learning, graph functions, path-and-pattern and text functions and gives them to the users in a way that is point and click. With that, you can select the algorithms and data elements that you want to run. Business analysts will be able to point and click, run these algorithms and find results without having to spend the time coding or going to their IT department or to a data scientist to ask for help. Now they can help themselves. They can access new sources of data.
Probably the most important functionality is iteration. Analytics, as we all know, is an iterative process. You ask a question and get an answer. The answer to that question spawns ten more questions that you want to ask. You want to iterate. Again, if you’re hand-coding, this becomes difficult, especially if you’re dependent on someone else. If you can point and click and, in a matter of seconds, produce a new outcome to your question, you can now iterate much faster. So by giving them this point-and-click interface with nice graphics, it will empower them to use machine-learning and advanced analytics to augment their decision-making process, make them faster and more efficient, and really help them get to the best outcomes. That is what Teradata is doing to democratize data science, machine learning and advanced analytics.
Those are some important steps forward toward democratizing data science.
Imad Birouty: They are. I’ll just add one more thing. I’ve heard some folks ask if Teradata is getting out of its sweet spot as a data warehouse company. They wonder if this is different for us.
Is this a little out of scope, maybe, for Teradata?
Imad Birouty: I say it’s actually very much in scope. Teradata started 40 years ago. Teradata today is very different than Teradata was 40 years ago. In those early days of relational databases, we had rows columns, very simple aggregations, very simple joins and so on. Throughout the years, Teradata has continued to add newer capabilities and technologies in geospatial, in-memory, time series and other math and statistical functions. Teradata has always been about the ability to analyze data and throughout the 40 years has continually added more functionality to make the analysis of data richer and easier to use. This is just one additional step in that direction.
It does make sense when you think about it. It’s very exciting, and we look forward to seeing what these newly empowered knowledge workers do with their data science capabilities.
About the Author
Phil Bowermaster is an independent analyst and consultant specializing in big data, business intelligence and analytics. Phil is the founder of Speculist Media, which produces blogs, podcasts, and other social and traditional media exploring the role of technology, particularly data technology, in shaping the future. He works with select clients in developing and executing content strategies related to big data. Phil can be reached at firstname.lastname@example.org.
All Rights Reserved. Copyright 2004 — 2019, TechTarget BeyeNETWORKTM, a TechTarget company