While prior educational qualifications in relevant fields helps in becoming a data scientist; it is not a mandate in any way. For becoming a data scientist, a candidate must be aware of the basics of statistics and mathematical algorithms, be aware of the basics of software engineering, and be self-motivated and patient to learn new tools, technologies and evolving methodologies & algorithms to implement them on real-world data.
An under-graduate degree in mathematics, statistics or even computer science & engineering could be a brilliant starting point; post which candidates need to learn scripting languages and statistical algorithms to take it forward.
The “Data Science Space” is just starting to heat up. It has been existing in various forms such as data warehousing, computational analytics and high performance computing in past. Only recently is it getting a lot of press.
I cannot speak for the whole industry but on our view there is a lot of education that needs to be done in the market. Companies that help other companies make sense of data right away typically win. So a good combination of good ETL skills, database and visualization typically helps.