Data science is a highly specialised field. As a result, tasks such as data cleaning and data preparation become critical for organisations as any inaccuracies may lead to project failures. The risk is even higher when non-specialists handle the data. “To build truly data-driven organisations, we need to unlock the power of data for as many people as possible,” said Francois Ajenstat, Chief Product Officer at Tableau.
Tableau recently introduced Tableau Business Science, a new AI-powered analytics class to lower the barrier of entry to data science, enabling business owners and analysts to make smarter decisions faster — with all clicks and no code. It has integrated with Einstein Discovery to provide real-time predictions and recommendations to help them explore business outcomes.
We got in touch with Francois Ajenstat to get insights into Tableau’s unique offering for the data science community. Ajenstat is responsible for Tableau’s overall product strategy and oversees the product portfolio of the company.
AIM: Data scientists spend a lot of time cleaning and preparing data. What are some of the challenges here?
Francois: Data scientists and analysts, in general, spend 80% of their time in data preparation. The reason is that we have more data than ever before, and the ability to work with data is still challenging. It requires deep data knowledge and specialised skills, such as coding, to make it work. To overcome this challenge, Tableau embarked upon the journey of self-service data preparation years ago to make data preparation somewhat simpler and intuitive for business users. Instead of building it from a code standpoint, we took a visual approach. We believed that it would be easier to clean the data if they see that data — a product that we called Tableau Prep Builder.
It now serves as the foundation on which we are adding new capabilities. Business Science, our new capability, is built on this premise, where it eases the task of data preparation while making the task of building models easier. It connects data preparation, data analysis and validation — all in one platform.
AIM: Tell us about Tableau Business science. How will it help the data science and business community?
Francois: In data science, working with data and building models can be quite challenging. We need data expertise to carry out these tasks, the lack of which leads to very few projects being done. Moreover, a lot of these projects end up failing. But hiring specialised resources can be pretty expensive.
When we think about this problem, there is an opportunity for us to deal with it and develop techniques that can be valuable for our customers. We are in the process of democratising data. We started with data preparation and are now introducing capabilities that we call Business Science. The task of Business Science is to use AI, analytics and data science techniques and provide tools and advanced capabilities to the business users who are not necessarily programmers. They know their data, but they do not necessarily have data science skills. With Business Science, we enable them to make decisions faster and use data science to drive better outcomes.
AIM: How is Business Science different from data science?
Francois: It’s the difference between being extremely precise vs getting quick answers. There is always a trade-off between the two. Tableau Business Science is more directional. We do not have to be 100% precise, but we have to be directional. We think that there is a need for both. Think of Data Science and Business Science giving you directions. While Data Science will address latitude and longitude, Business Science will tell you to take the next left or next right. That is the difference. It is about precision but takes you in the same direction.
With this concept, we have used the Einstein Discovery platform by Salesforce that brings analytics and machine learning capabilities to business users. We are using this platform to drive millions of predictions every day to Tableau users. It will work on more types of data sources and types of users, leveraging the ease of use by Tableau.
AIM: How will Tableau’s innovation help organisations drive business insights with predictive analytics, which usually requires a specialist?
Francois: Historically, Tableau has focused more on diagnostic analysis — what happens and why did it happen. With this new capability, we are doing it in the predictive and prescriptive analytics way. You can see Einstein’s prediction, the reasons and factors, and why it is driving that prediction. It will also show prescriptive insights about what you should do about it. We are enabling this capability by sharing a fully interactive dashboard. You can also enrich datasets with predictive insights.
For example, in the shipping industry, it can predict the likelihood of the product shipped from a particular country to reach on time. All the predictions are interactive and provide the ease of asking questions and derive results with a click. It can be used by a wide range of industries such as finance, e-commerce, and more since it works with clicks and not codes; it provides business users with ease of use.
AIM: Why has data democratisation become an essential requirement for businesses today?
Francois: Data democratisation is a critical requirement today as it provides access to data. Democratising data science will help more people make smarter decisions faster. There is a difference between thriving and barely surviving. Now that we are all digital and work from anywhere in the world, the speed of digital transformation has accelerated dramatically. Digital transformation is a data transformation. As customers engage through digital means instead of physically, it exposes them to dramatically more data to bring to their advantage to design better marketing campaigns, better customer service or better engagement overall. We believe that customers that transform digitally also transform from a data perspective. Technology alone does not drive success. It requires organisations to have a data culture, i.e. thinking of data as a critical element of their success. It is also important that people have access to self-service with governance.
AIM: What are the challenges data scientists and business users face today?
Francois: One of the most challenging aspects is communication. It’s one thing to get insights, but you need to share your insights with others. At Tableau, we aim to make data understandable. We believe in making people data literate and increase the data communication skills of the people. Especially in today’s environment, we need to have speed and agility, and Tableau provides that. Throughout the years, we have been known for providing ease of use of tools and being approachable, as a result of which people can think with data and deliver projects faster and with more efficiency.
AIM: How to make sure data analysis gets used to its full potential?
Francois: There are always a couple of different aspects — the first being that there is a lot of data but a lack of insights. Most users are collecting large volumes of data but are not able to make sense of it. The second aspect is that not all data have a shelf life. Sometimes users need quick answers. In such cases, decisions have to be made based on shelf life to make sense of it. For that to happen, things need to move quickly while getting the right kind of insights. Solving these two challenges will lead to better use of the analysed data.
AIM: What is the roadmap and growth plans of Tableau for 2021?
Francois: This year, we are building a lot of capabilities around Tableau Business Science. It is the beginning of a big investment that we have made around democratising data science. Another big focus will be on data management, i.e., enabling people to catalogue, curate, clean and prepare their data for analysis. We made many big investments in data management last year, and we are continuing on that journey to make data management easier and accessible to more people. There will also be a lot of investment around engagement to help people collaborate with data.