Follow Datanami:
September 20, 2021

An Interactive Analytics Whiteboard for COVID Times

(Source: Einblick)

COVID has been a boon for remote work and technologies that support it. Zoom, for instance, has more than doubled its stock price. But when it comes to enabling data analysts and data scientists to work together in real time, the technology enhancements have not been as forthcoming. Now a Massachusetts company called Einblick is aiming to change that.

Einblick develops an interactive analytics whiteboard that allows users to visually explore their data, build models, and make data-driven decisions as a group. Based on six years of research at MIT and Brown University, the platform helps users overcome the challenges of remote collaboration.

“We see three limiting factors which hinder collaboration on data problems,” Einblick CEO and co-founder Tim Kraska tells Datanami via email. First and foremost, data science GUIs are typically designed to be used by a single user at a time. “They simply do not support several people to work on the same screen at the same time,” he says.

Secondly, for complex operations or big data, results are not created at interactive speed, which destroys collaboration. Getting results back quicker–even if they’re just approximate results—helps the collaborative process move forward.

Lastly, there are just too many tools involved, Kraska says. “One tool for video conferencing, another one for data cleaning, another one for model building, etc.,” he says. “This makes it hard for a team to stay focused on the problem as it requires constant context switching and exporting/importing between tools.”

By comparison, Einblick presents a centralized offering that includes a full stack of the tools and technologies needed for analytics workflows, starting with data cleansing and transformation and ending with model building and what-if analysis.

The Einblick analytics whiteboard allows teams of users to collaborate on a single data set (Image source: Einblick)

The heart of the product is the data whiteboard. “It combines the fast data exploration and visualization capabilities of tools like Tableau, with the capability to create complex workflows, like in Alteryx,” Kraska says. The company calls this concept “visual data computing.”

The offering makes use of what it calls a “progressive approximation engine” that computes results on small samples of data. The computation continues as sample sizes increase, until eventually it converges on a final result over the entire dataset, the company says. This sacrifices some initial accuracy for speed, which is essential in a collaborative setting.

Today, the Cambridge, Massachusetts company added a new element to the platform: a video-based collaboration feature designed to make collaboration more natural among distributed users. Now users can have a Zoom-like call while collaboratively working with data and testing different hypothesis. It also announced that free 14-day trials to groups that want to take Einblick 1.0 for a test drive.

While the Einblick platform includes a range of tools and tech, the company is not trying to replace the complete analytics stack. Rather, it’s trying to bring small semblance of real-time interactive analytics when users are collaborating. That is an area of need, particularly as workers have been forced to work from home during the COVID pandemic.

“We believe deeply that most organizations need a way to build simple models to inform their business users,” Kraska says. “Creating a product that allows teams to move from data transformation to visualization to model building all in a single interface is more important than optimizing each of those steps in a dedicated vendor system.”

To some degree, you can consider Einblick as a “code-last Jupyter notebook for teams,” Kraska says. “However, internally it works very differently. Jupyter Notebooks are designed for serial editing with one flow (top to down), whereas Einblick is an infinite canvas where teams can have several ongoing threads and ideas explored at the same time.”

Einblick is also open to integrate with other tools and libraries as new visual operators. It also supports exporting to standard Python, such as those used with Juypter and other data science notebooks, and importing data using standard formats.

“We find our best customers use Einblick as the place where technical (data scientists) and non-technical (business SMEs) employees meet to explore data related to a business problem,” Kraska says. “In addition, we aim to enable more business users to be able to solve more and more complex use cases by themselves.”

Einblick is offered as a cloud solution. There is also an on-prem version avabible. For more information, check out www.einblick.com

Related Items:

Still Wanted: (Much) Better COVID Data

Open Data Hub: A Meta Project for AI/ML Work

Databricks Brings Data Science, Engineering Together with New Workspace

 

Datanami