Technology companies are launching data science platforms of widely varying complexity. Some claim their tools are so simple they will make “citizen data scientists” out of folks with little or no experience. TheCUBE, from the SiliconANGLE Media team, recently spoke with an actual research scientist who offered a caveat for everyone working with data: Before you jump up and down about the new model you made, check your Extract-Transform-Load.

Miriam Friedel, research scientist at Elder Research Inc., spoke to Sam Kahane (@Sam_Kahane), during the IBM DataFirst Launch event about the current state of data science.

Having just attended a data science panel, Friedel remarked on some universal snafus practitioners encounter. “It was really interesting to hear about some of the common challenges that we all face working on different types of data science problems,” she said.

Building data models that walk the walk

“One thing that we struggle with a lot is it’s not simply enough to build a model that’s really interesting or technically advanced,” she explained. “You need to put it into production, deploy it, and that’s what brings value to your customers. And a lot of other people on the panel had spoken about similar things.”

Keeping one eye on ETL

Data science is intricate — and those working with it — be they Ph.D.s or small business owners — need to heed this to be successful, Friedel advised.

“If you happen to be building a model and it looks too good to be true, it probably is, so make sure you check your ETL and your modeling that you’re not doing something crazy,” Friedel concluded.

Watch the complete video interview below, and be sure to check out more of SiliconANGLE and theCUBE’s coverage of the IBM DataFirst Launch event.

Photo by SiliconANGLE