In the wake of the recent presidential election, it’s obvious to many more people that a fair number of flaws can arise when the models used to construct predictions using Big Data analytics are not especially transparent.
Aiming to solve that issue, Domino Data revealed today it has generated an additional $10.5 million funding to drive development of a data science platform that makes it possible to collaboratively create data models that can be more easily vetted by others.
Domino Data CEO Nick Elprin says one of the inherent flaws associated with data science is that interpreting the data is still too much of an art practiced by one individual rather than a team. That makes it hard to identify the biases that went into building the model. Without that kind of transparency, business leaders become concerned about making decisions based on interpretations of the data that wind up being deeply flawed, says Elprin.
The goal, says Elprin, is to make it possible to build data models faster, but also fundamentally better, because more data scientists can provide input.https://o1.qnsr.com/log/p.gif?;n=203;c=204663295;s=11915;x=7936;f=201904081034270;u=j;z=TIMESTAMP;a=20410779;e=i
“There’s a clear need for more transparency into the data models,” says Elprin.
Obviously, after the elections, many business leaders are wondering if all the investments they are making in Big Data will be worth the time and effort. After all, if analytics tools could be trusted, there would be a different President-Elect than there is right now. To regain the confidence of those business leaders, it’s clear that data scientists will be required to plainly show how the conclusions they are drawing are being reached.