There's no doubt that data science is a hot topic right now and that data scientists are in high demand. The prospect for an analyst to become a data scientist is quite compelling: great pay, exciting work, and the opportunity to work for some of the best companies in the world. That said, most of the analysts that I encounter who are interested in pursuing a career in data science, aren't really sure where to start. In all actuality, it's not hard for an analyst to become a data scientist, but you must be open to learning a few things and putting a good plan in place.
Data science is a multidisciplinary skill. If you look at the field of cognitive science--which is tangential to data science--it's a fascinating conglomeration of psychology, philosophy, artificial intelligence, anthropology, and neuroscience. In the same way, data science brings together a synergistic array of skills including: computer science, machine learning, advanced mathematics, and data visualization. I've seen data scientists emerge from each of these corners of the field. For instance, I've seen quite a few physicists augment their talents and skills in advanced mathematics with computer programming. I've even seen graphic artists, who realize there's gold in the hills of data visualization, build a supplemental education in math and computer programming.
So, the first step in becoming a data scientist is to do a self-assessment around these four areas: computer science, machine learning, advanced mathematics, and data visualization. Where are your strengths right now? Where do you need to improve? Once you know where you stand, take it upon yourself to build competencies in your weak areas. If you're weak in machine learning but have a good math background, pick up a few books on the subject and find out what all the hype is about. If you're strong in computer programming but haven't looked at statistics since college, take a refresher course online. The amount of freely available education on these subjects, like those offered by the Khan Academy, is truly remarkable. It's better to become a data science enthusiast / hobbyist before you pursue a job as a data scientist. It's very unlikely that a company will hire you as a data scientist and then allow you to learn on the job.
Your strategy for transitioning from analyst to data scientist should be both proactive and reactive. It takes more than just skill and talent to become a data scientist--it takes a plan. While you're building your competencies, you should be building your brand as a data science enthusiast. Take every opportunity to let people know that data science is your thing. Join associations, participate in discussions online, and volunteer for efforts within the company that will help you practice your skills and build your brand as a data scientist.
You should also consider a formal education in a field related to data science. For instance, the UC Berkley School of Information now offers a Master of Information and Data Science program. This is just one of many top universities that are offering degree programs appropriate for a data scientist--many of which are online. Not only will this fortify your knowledge of data science, but it will also provide the necessary weight in your resume to actually get hired as a data scientist. Most companies won't hire you as a data scientist unless you have an advanced degree in the field.
Finally, look for opportunities within your current job to move into a data science role--this is where the reactive part comes in. You must always have your antenna up for projects, teams, and job positions that involve data science. Companies today are still experimenting with data science, so it's not like Marketing where there's a clear career path. There are usually small windows of opportunity to plug into data science efforts, so you must we ready to jump when the time comes. Having both the education and some experience on your resume will position you properly for your next job as a data scientist.
Becoming a data scientist is an exciting proposition for an analyst. Not only is the work challenging, but since data science is in such high demand right now, the pay is very good. The transition from analyst to data scientists isn't difficult, but it takes some work. First, build your own competencies in computer programming, machine learning, advanced mathematics, and data visualization. You can start today by taking an online course and/or enrolling into an online degree program. Then plot a course for your success including branding, education, and experience when the opportunity presents itself. It may take some work, but eventually top companies will be calling you to be their next data scientist.