A startup called Datahero launched on Thursday with a new cloud service that makes visualizing data as simple as a few mouse clicks. The company has raised $1 million in seed funding led by Foundry Group, as well as individual investors that include the founders of Aster Data Systems, former MarkLogic CEO and Salesforce.com SVP Dave Kellogg, and data scientist Mike Greenfield. Datahero is just the latest in a string of companies trying to simplify the process of data analysis and visualization (Tableau Public, Platfora, and ClearStory spring to mind as some others), but it also targets the broadest audience — you and me.
Datahero’s mission, Co-Founder and CEO Chris Neumann told me, is to let “literally any user … ask business questions and get information [by themselves].” It’s an important goal as we move toward a future where data will play an ever-more important role in decisions both business and personal. Not everyone can be an expert data analyst or data scientist, but everyone will need some sort of analytics tool at their disposal (GigaOM Pro subscription req’d).
In order to build a product product capable of fulfilling that mission, Neumann, an early engineer at Aster Data Systems, teamed with Co-Founder and Chief Product Officer Jeff Zabel, who came from BMW where he was in charge of designing user interfaces for third-party Internet-connected apps such as Pandora. The two think their mix of data and design expertise will make Datahero a big hit among users more familiar with consumer services such as Mint.com than with business intelligence applications.
I had a demo of the product, and it looks like Neumann and Zabel are onto something. Dragging and dropping a CSV file into the browser (or linking to an online data source) to get started is easy enough, as is the process of creating a workable dataset from it. Datahero tried its best to characterize proper categories and data types from the fields, but even when it messes up it’s easy to fix using a pulldown menu. Neumann said they learned from Mint.com that consumers don’t mind fixing errors, especially when an app is quite accurate to begin.
“We’re going to make it easy for the user to help us,” he said. Neumann’s big data background doesn’t hurt the process either. He said the product uses machine learning techniques to get smarter about characterizing data types with each new user.
Once the dataset is done, it’s just a matter of dragging and dropping values into the canvas area to create a chart. Datahero will generate the chart it thinks is best suited for your data. A collection of call records, for example, automatically generate a map showing the frequency with which particular area codes were dialed. The product even tells users what fields they should or shouldn’t add as they try to drill down deeper with their data, to avoid creating a messy or non-sensical graph.
To me, the freemium Datahero functions a lot like Tableau Public, only easier for the layperson who understands very little about X and Y axes, proper data formatting and analytic functions. For users who don’t even want to lift a finger to drag and drop fields, for example, Datahero offers predetermined graph types based on the type of data a user has and what other users have made from that type of data in the past. Although there are capabilities for users with more analytics knowledge, Zabel said the goal is to “make the average user a hero” and only “advance the advanced user.”
However, if this all sounds just a little simplistic given Neumann’s background and massive unstructured datasets and MapReduce at Aster Data, it’s only temporary. Although he and Zabel are keeping mum on the exact plans, Datahero will likely start to look something like Google Apps with regard to specific features for basic, advanced and enterprise users, and complex analytics (while still maintaining the consumer user experience) are coming. The company’s investors wouldn’t have been so excited, Neumann said, “if this was just another graphing layer for Excel.”
Bonus material for the developers out there: Datahero’s application layer is written entirely in Node.js and is hosted on the Joyent Cloud. You can learn a lot more about Node.js at our Structure conference in a few weeks, where we will be speaking with Joyent’s Jason Hoffman on the topic.