Once done, the individual needs to bid for projects which may seem doable and submit the proposal for the same. The bid would then compete against thousand other bids; and then a single bid would be selected by the project sponsor.
Of course. In fact, a lot of famous data scientists solely work as freelancers. I personally too have freelanced for a European Retail based client. Make sure you have a good profile on GitHub, Kaggle where the result and the code are optimized and not just generic as others. You should register yourself on a freelancing website. However some prior experience always helps in getting clients. Toptal is one of the best freelancing websites, so if you clear their exam and can get into the community then nothing like that.
In fact, consulting companies will be happier to hire a freelancer over regular employee as it will allow them to optimize the costs in the periods when no projects are running and not much workforce is needed.
While there is a shortage of data scientists, universities are ramping up undergraduate degrees programs in data science. Based on the current job market that is about one-tenth of the computer science market, there will be many college graduates to fill many of the existing openings at a lower cost.
Most data science jobs deal with confidential information requiring business analysis, and these jobs are rarely farmed out. The job market is really in business analysis that requires more than just data science.
I believe there are too many dynamics to say if you can become a data science consultant. It may be cheaper to hire than to outsource data science jobs.