Companies are learning how to collect all kinds of data, but they aren't learning how to do data science. That was the message from Kim Stedman, former analytics manager for the Seattle-based gaming company Meteor Entertainment. She recently gave an Ignite talk, a storytelling event for geeks sponsored by O'Reilly Media, where she called out data scientists everywhere: It's time to step up and take control of the data science terminology sprawl.
"[Companies] are looking for everything; they have no idea what they need," Stedman said.
Yes, companies have bought into the buzz and are willing to dole out big bucks to bring data science talent in-house, but ask them to outline the specific skills they require and it's more conceptual art than scientific blueprint. That lack of clarity doesn't help things much when a data science team, which can be a costly expense, is expected to produce immediate value and doesn't, Stedman said. Businesses may want the rigor of the academy, but they're rarely willing to pay for the time good science requires.
"Science is the assertion of a testable hypothesis followed by a cycle of validation and refinement in pursuit of a working theory. It challenges beliefs," Stedman said. That kind of new world order doesn't produce rich results overnight. "Data is fast, but science is slow."
One ring to rule all
A campaign on Kickstarter -- the platform that crowd-sources fundraising efforts for independent projects -- raised more than $300,000 to build near field communication (NFC) technology into jewelry. The campaign, launched by John McLear out of the United Kingdom, is specifically focused on producing NFC rings, which will be able to unlock smartphones, doors and even link people together when the ring touches another device. Two separate transmitters built into the ring separate private data from public data.
Know what the biggest prime number is? Try 2 to the 57,885,161th power minus 1. It was discovered earlier this year by Dr. Curtis Cooper, a mathematician currently teaching at the University of Central Missouri, and was the culminating moment in Adam Spencer's recent TED Talk. The number is almost 17.5 million digits long.
"If you type it on paper and saved it as a text file, that's 22 [megabytes]," said Spencer, a comedian who happens to also love math.
Data begets data
Acxiom Corp. made the news this week with its site aboutthedata.com -- but it may not be the kind of news the data broker wanted. The portal allows visitors to find out the kinds of data Acxiom has collected on them for its corporate clients. But here's the kicker: In order to see the results, you have to -- what else? -- cough up your data. The login asksfor your name, address (physical and digital), birthdate and the last four digits of your Social Security number before granting access to the data they've collected through other means. As critics pointed out, so much for data transparency.
Magic Quadrant for mobile app dev platforms
Gartner Inc. released its Magic Quadrant for Mobile Application Development Platforms (MADP). SAP, jQuery Mobile, IBM, Kony, Adobe and Antenna were given the highest marks, followed by MicroStrategy and Salesforce.
The report also acknowledged the difficulty in choosing an MADP in today's market. It provided CIOs with several words of caution, including:
- Native vs. Web vs. hybrid apps is still up for debate. And, although HTML5 is full of promise, "most enterprises find it is only a partial solution today," the report stated.
- Finding one vendor for the entire organization isn't likely, at least not yet. The needs of the organization overwhelm what vendors can offer, at least right now.
- The market is on fire with "new vendors and new products entering the market every month. At the same time, the market "continues to converge," making one vendor from the next almost indistinguishable.
Claudia Imhoff, president and founder of the Boulder, Colo.-based consultancy Intelligent Solutions, broke down business intelligence (BI) during a recent webinar with The Data Warehousing Institute.
1. Descriptive BI: The most prevalent form of BI and also the least valuable, she said. It describes what is happening for well-defined business problems or well-defined opportunities and relies on business reporting, dashboards, scorecards and data warehousing.
2. Predictive BI: A more proactive approach that tries to answer the questions, "What will happen and why." The technologies here are different from descriptive BI and more complex, including data mining, market basket analysis, text mining and forecasting.
3. Prescriptive BI: The ultimate goal of BI, the goal here is to identify what are the next best steps to take and why, she said. With this form of BI companies have to become adept at doing things like simulation and decision modeling.
Semi-structured data is king of LinkedIn analytics
Community cloud could fix data crunching dilemma for cancer research
Social collaboration software through the lens of a SharePoint lover
Here's how to hire data scientists
Get advice from the experts on how to build a data science team for your business