New data models for a new world

Nearly all existing models have been invalidated by the coronavirus pandemic and will presumably be invalidated again as the situation evolves and as the world recovers.

Even after recovery, it is unlikely that we will return to where we started before the crisis began.

In this time of unprecedented disruption, data scientists need new models that meet changing trends and new consumer demands. Automatic Feature Engineering is one of the top methods of achieving this goal. By uncovering novel patterns buried in complex data, data models can perform better and generate timely, accurate insights that companies need to grow and thrive.

SparkBeyond is uniquely positioned to help businesses adapt to – and even thrive in – this New Normal. By using groundbreaking AI-powered technology, companies can uncover new insights and generate new ideas in a post-coronavirus world.

Old data vs. new data

In the span of a few short months, the world has changed. As Data Scientists have seen a major shift in their focus, there is now a clear need to understand and separate old data (before COVID-19) from new data (after COVID-19). It’s critical to learn what characterizes this data and understand new drivers that are appearing in this new normal.

“Most of the risk models we have not been able to capture the shift in the landscape”, noted a former Head of Investment Solutions and Products at a global bank. Up to 80% of data used in existing models require considerable modifications to reflect changes and identify new patterns which are still emerging. This thinking permeates nearly every sector of business today, and it requires solutions that can uncover new patterns and create new insights quickly and accurately.

The ability to correlate between old and new data – even as new data is limited or sporadic – is a critical element of AI-driven business solutions.

"Going far beyond analytics"

More than simply providing in-depth analyses on existing data, SparkBeyond augments with a rich network of external data sources and offers an in-depth perspective of challenges facing our society.

The AI engine is infused with a creative thinking algorithm which allows it to connect the dots and think out of the box as it tests 4 million hypotheses per minute on complex data. This leads to the discovery of novel drivers, root causes, microsegments, and features that a domain expert would never think to test or a human eye could ever spot.

By merging client data with external information such as geospatial data or footfall traffic, data scientists obtain a more holistic view. They can identify detailed and complex behaviors that simply didn’t exist even just a few weeks ago, such as new spending patterns or how lockdown restrictions impact businesses in multiple regions.

The platform increases accuracy by training models from early-impacted areas. Data scientists can use this to create new forecast models with the most recent data, while the engine automatically rebuilds the models’ core on a weekly basis.

Separating signals from noise

Even with newly acquired data, it must consistently refresh itself to provide analysts with relevant and effective outcomes. The SparkBeyond platform is configured to scan for new patterns at specific frequencies (e.g. daily, weekly) and calibrated to separate signal from noise.

The ability to drill down into a variety of features, pinpoint high lift micro segments, and focus on a particular subpopulation, such as only people who churned or only companies who spent more than $___, provides high level insights. These allow companies to make smarter decisions for their customers, identify new markets, and create new products and services.

Recent use cases

What are the key drivers of new shopping habits and consumer purchasing behaviors?

As home isolation policies around the world have gone into effect, consumers are forced to change the way they shop. For example, since bars and clubs have closed, the sale of draft beers has decreased significantly. However, beer sold in cans and bottles has increased, as they can be purchased in supermarkets. Consumers with a history of purchasing mid-level brands are now showing a preference for more expensive brands, uncovering new sales and marketing channels for beverage companies.

The hospitality industry has experienced dramatic changes as air travel has declined and social distancing measures have begun. At the same time, we’re seeing positive changes for rental property and hotel room bookings. As a recent use case, the SparkBeyond platform discovered that including the word ‘disinfect’ in room descriptions increased the likelihood of a booking. As hygiene has become a top priority in the new normal, it’s also become a sales driver for the hospitality market.


Features

No items found.
No items found.

"Empowering data science teams"

For data scientists to remain valuable partners to their business counterparts, they must be able to provide actionable business insights and useful predictive models in much faster cycles — to power the constantly-changing conditions that their businesses face.

Capturing, analyzing, learning, and testing ultimately allows Data Science teams to arm their companies with explainable insights and actionable solutions. Within just a few hours, the SparkBeyond platform can be deployed so you and your team can be up and running with a rapid training program. Leverage capabilities in just the first days of deployment and experience the power of next-generation AI technology in the new normal.

It was easier in this project since we used this outpout

Business Insights

Apply key dataset transformations through no/low-code workflows to clean, prep, and scope your datasets as needed for analysis

Predictive Models

Apply key dataset transformations through no/low-code workflows to clean, prep, and scope your datasets as needed for analysis

Micro-Segments

Apply key dataset transformations through no/low-code workflows to clean, prep, and scope your datasets as needed for analysis

Features For
External Models

Apply key dataset transformations through no/low-code workflows to clean, prep, and scope your datasets as needed for analysis

Business Automation
Rules

Apply key dataset transformations through no/low-code workflows to clean, prep, and scope your datasets as needed for analysis

Root-Cause
Analysis

Apply key dataset transformations through no/low-code workflows to clean, prep, and scope your datasets as needed for analysis

Join our event about this topic today.

Learn all about the SparkBeyond mission and product vision.

RVSP
Arrow