The obstacles data scientists face
Data scientists spend a lot of their time extracting and cleaning information before they can analyze it, and often this is harder than it has to be. One of the biggest obstacles in data science today is the siloed nature of information at a lot of organizations, reports Ellis Booker at InformationWeek.
Some companies put up their own organizational and infrastructural impediments to data analytics by retaining walled-off data sources fiercely protected by internal fiefdoms. Others maintain siloed data because of compliance requirements that prevent information-sharing, as in the highly regulated healthcare field.
At Salesforce.com, there are six data scientists and three business analysts on the product intelligence team. The team analyzes data from customer calls, sales visits, product research, website activity, billing information and social media to better understand customer sentiment and behavior. One of the projects underway right now analyzes activity on the company's customer-facing website called Idea Exchange, where members can post requests for new features. There are 26,000 ideas there right now. After the data science team analyzes the data, it send its conclusions to the relevant product team.
- see Ellis Booker's article at InformationWeek