T-Mobile asks job applicants to take this test before inviting them for an interview because the company has found powerful correlations between the online assessments and success on the job. High scorers tend to resolve customer calls about 25 seconds faster than those who receive low scores. That means they can handle one more call a day and about 250 more a year. [Via Tyler Cowen]
Most of a data scientist’s time is spent creating predictive models: finding the variables that matter to make predictions, the right type of model, the best set of parameters, etc. Work is being done to automate all of this, and so far it has resulted in solutions such as Emerald Logic’s FACET and in the creation of prediction APIs such as Google’s and Ersatz Labs’. These APIs abstract away the complexities of learning models from data. You can just focus on preparing the data (collecting/enriching/cleaning it), you then send that data to the API, it automatically creates a model, and it uses that model when you ask for predictions.
No. Most of a data scientist’s time is spent understanding business problems to solve, collecting and preparing data that can solve the problems, and later communicating results. Predictive modeling takes only a small portion of the time and it is not the only kind of analysis work that is done.
If your job is building models, all you do is try to build models. A data scientist’s job should be to assist the business using data regardless of whether that’s through predictive modeling, simulation, optimization modeling, data mining, visualization, or summary reporting.