Your Data Science Talent Is Hiding in Plain Sight

Dataiku Product, Scaling AI Dan Darnell

In a prior blog, I talked about how to enable expert data scientists to do more expert data science. While that is a great start, scaling AI to deliver tens of thousands of AI-enabled processes and applications will take more people in addition to some clever technology choices. 

Every business has people with technical skills hiding in plain sight that you can use to scale AI programs — they have titles like engineer, actuary, analyst, or statistician. Or, they could be like me; I have an engineering degree and technical aptitude but chose to pursue a career in business. These people are in functional departments like finance, talent, and marketing. Whatever the case and their title, you have people in your company with the skills to take on data science challenges, provided they have the right tools.

colleagues at computer

Dataiku is the platform for Everyday AI, with a visual, low, or no-code environment that enables a broad range of skill sets to take on advanced analytics projects. In Dataiku 11, we have further enhanced our offering for this skilled group to go even deeper into the world of data science, so they provide more value to business teams.

Visual Time Series Forecasting

Classical ML techniques struggle when data changes over time. Sometimes this change, however, is part of the signal, like seasonality in shopping trends, impacts of weather, and other repeated events. Incorporating these time-based events is called time series forecasting, which has typically been the domain of expert data scientists.

Time series forecasting has broad applications across industries and functions. Finance teams use forecasting for revenue and cash, sales teams forecast which deals will close, manufacturing engineers forecast maintenance needs, and more. Ideally, the experts within functional areas should be the ones to develop and maintain these projects because they have business context and are closest to the data and the problems that need to be solved.

Dataiku 11 enables time series forecasting as part of Dataiku's visual machine learning. The new capabilities allow users to design, train, evaluate, and deploy time series forecasting models without writing code or using notebooks. With this functionality, a broad range of users across the company can now build and maintain their forecasts using Dataiku. 

→ Learn More About Time Series Forecasting in Dataiku 11

Outcome Optimization

For business experts who understand the data and unique issues for their area, statistics that explain model functions are less meaningful than seeing how a model responds to inputs. This is why interactive what-if analysis is vital for these users to interrogate models by changing inputs and viewing outputs. Once users start changing values, they quickly begin looking for an optimal outcome based on their knowledge. Outcome optimization in Dataiku 11 enables users to quickly and visually see which combination of values generates the best results. With this new information, business users can look for ways to change and influence processes to drive more desirable outcomes.

→ Learn More About Outcome Optimization in Dataiku 11

Seamless Sharing 

One of the most significant barriers to scaling AI is the manual creation and re-creation of objects like datasets, models, reports, and applications. Projects in one group can be similar or even the same as others, yet each group creates their project from scratch, potentially wasting time and money. Ideally, organizations want to create and share approved assets and projects for teams to use and learn from. Enhanced sharing and access workflows in Dataiku 11 enable groups and organizations to share assets to maximize the use of approved assets and minimize rework which benefits everyone and helps scale AI and control costs.

→ Learn More About Seamless Sharing in Dataiku 11

Visual Logic

When business users look at a dataset, they often see an opportunity to organize and augment the data with rules-based logic based on their experience. Creating this logic can become complex with multiple nested if-then statements, which can be challenging to debug and even more difficult to explain. Rules are also valuable for many other project areas like filters, joins, and reports, but the same risks apply. New visual logic capabilities in Dataiku 11 simplify complex rules development with new functions throughout the interface. With these new capabilities, business users can more easily tailor data, projects, and outputs based on their business insights to drive more valuable outcomes for their teams and organization.

→ Learn More About the New Visual Logic Capabilities in Dataiku 11

To meet the demands in our organizations for AI-driven processes and applications, organizations must enable a new set of technically minded users across business functions. With Dataiku 11, business teams have even more visual capabilities to take on new challenges like forecasting and optimization that can drive value for their teams and companies.

You May Also Like

Why Is Data Architecture Important for the Success of AI?

Read More

4 Trends for AI in Life Sciences for 2023

Read More

3 Steps to Win Pole Position in the AI Race

Read More

4 Trends for AI Financial Services in 2023

Read More