This post was originally published on this site.
Key Takeaways
- Focus on Learning Python: Start with mastering one programming language, like Python, and simultaneously learn statistics, data wrangling, data visualization, and machine learning fundamentals.
- Leverage Transferable Skills: When moving from a different industry, highlight transferable skills and reword your experiences using data science terminology, matching job descriptions.
- Develop a Structured Learning Plan: Create a learning plan based on job descriptions and bootcamp syllabi to identify and fill knowledge gaps in data science skills.
- Prioritize Portfolio Projects Over Formal Education: Build a portfolio with diverse projects to demonstrate skills and hands-on experience, which are often valued more than formal education by employers.
- Maintain an Online Professional Portfolio: Use platforms like GitHub for project details, while summarizing and discussing projects on Medium or LinkedIn to enhance visibility and showcase problem-solving capabilities.
- Stay Informed and Continuously Learn: Regular reading of books, blogs, and academic journals is crucial for staying updated with data science trends and technologies, distinguishing between mid-level and senior-level data scientists.
Transitioning to a Data Science Career
Navigating the shift to a data science career from a non-technical background can seem overwhelming, with challenges like mastering new skills and translating diverse experiences into the field. Adam Ross Nelson, a seasoned data scientist and mentor at Adam Ross Nelson Coaching, breaks down this transition into manageable steps in our upcoming Q&A. He offers practical advice on essential programming skills, leveraging transferable skills, and building a standout portfolio. Dive in as Adam provides key strategies for confidently stepping into the data science arena, emphasizing the importance of hands-on experience and continuous learning.
What are the most critical skills someone coming from a non-technical background needs to develop to break into data science?
“My nutshell quick version of this is to pick one programming language starting out. Just one. Don’t worry about learning them all. Learn one language really well. Most will start with Python. While learning Python you can also learn statistics, data wrangling, data visualization, and machine learning fundamentals.”
Could you provide insights on how to effectively translate and apply skills from a different industry to a data science role?
“Two things come to mind. One, lean into those transferable skills. Two, do portfolio work and projects.
More on those skills though… when transitioning from a different industry to a data science role, it’s essential to identify and emphasize transferable skills. And there is more to this. You also have to learn to re-word your skills. You have to translate the words that you used to communicate your skills in the previous career to the words that data scientists use.
For example, a school teacher might say on a resume “taught classes in math and statistics.” On a data science resume that could be restated as “provided non-technical audiences with clear and concise explanations of highly technical concepts in math and statistics.” This is a rough guess on how it might sound to make that translation. The best thing to do is to find words and phrases on the job description that describe what you did in your previous roles. Use the terms and phrases from the job description in your resume.”
What are the key steps in creating a learning plan for someone transitioning into data science, especially for self-learners?
“My best advice on this is in two parts. First, look at job descriptions to identify the skills and abilities employers are looking for. When you see terms and phrases you don’t recognize write those down and then find learning resources that will fill those knowledge gaps.
Likewise and second, visit at least three or four major bootcamp program websites. Get the syllabi from those programs. See what they’re teaching. When you see terms, phrases, and topics you do not yet know, write them down, and then find learning resources that fill those knowledge gaps.”
How important is formal education in data science compared to hands-on experience, and what balance do you recommend for career changers?
“What even does “formal education” mean anymore? If you mean college and university education I caution you. Public confidence in the value of college and university education is falling, and some say for good reason.
The smartest employers know that skill is not equivalent to experience. Related skill is not equivalent to education. Skill comes from experience. This is good news for mid- and late-career professionals especially because folks who have been in the working world for more than a few years have more fodder to talk about when employers ask about hands-on experience. And this means mid- and late-career folk have more hands-on experience that they can reference and show through portfolio projects. Instead of formal education, portfolio projects have grown and will likely continue to grow in importance for folks transitioning into data science.
I’m also a strong believer in boot camp style programs. And if you’re not convinced that bootcamps are valuable alternatives to so-called “formal education” you’ll have to argue with major employers who are no offering internal boot camp programs in order to internally transition existing employees into data scientists.”
Building a Compelling Data Science Portfolio
What types of projects or work should be included in a data science portfolio to best demonstrate a wide range of skills?
“On this question I first like to define “portfolio.” Your professional portfolio is anything anyone interested in your skills can find about you online. In short, it is the first ten search results on Google when anyone searches your name or email address.
In my book I give ten portfolio project ideas. Here are three of them that are among my favorites. First one, collect original data, publish it, analyze it, and invite others to do the same. Second one, put a sentiment analysis project into production. With the pre-trained models from Google, AWS, Azure, Natural Language Toolkit, and other sources it is no more than a weekend project to put a sentiment or natural language processing project into production. Third, make a rosetta stone project. Making a rosetta stone project involves converting an existing project from one language to another language. Example, find a project you or someone else produced in the R programming language and then port it over to Python. Or the other way around (Python to R). You will learn a lot from this and it will show your programming skills very well.”
How can a data scientist best showcase their problem-solving process and results within their portfolio entries?
“I am a fan of disseminating projects via Medium and LinkedIn articles. The advantage of using platforms such as Medium and LinkedIn is that the articles you write will quickly rise high in search engine results. Often articles on Medium and LinkedIn will rank higher and more quickly than articles posted on a vanity URL portfolio website address.
While the meat and bones of your portfolio projects will likely reside on GitHub, it is important to write articles about your work too. The articles you write will link back to GitHub for those who have an interest in going deeper.
A good article that talks about a project might start with a description of the problem statement the project might solve. Followed by an outline of the methodology, and highlight the key steps in data collection, cleaning, and analysis. Descriptives of the data including proper visualizations help too. The article should also document any challenges and solutions. Finally. clearly present the results and discuss options for future work by yourself or others.”
Can you suggest strategies for those with limited professional experience to build a portfolio that stands out to potential employers?
“My advice on this is that less is more. Keep it all as simple as possible. Use the advice I mentioned a moment ago such as keeping the core of your project on GitHub while you also write about your projects on LinkedIn and on Medium.
However, for those that truly wish to stand out one of the best steps you can take is to ask trusted colleagues to read your work and offer feedback. Have your readers tell you what they think isn’t clear or what doesn’t fully make sense. Use that feedback to revise and improve your work.”
Career Advancement Strategies in Data Science
What are the most effective resources for a data scientist to stay updated with emerging trends and technologies in the field?
“In a word I say, “books.” Reading about data science is a total hack. If you’re like me, reading does not come easy. However a sustained and focused habit of reading a little bit each day about data science, Python, advanced analytics, and related topics can help you grow in how you understand and think about the field. The better your understand the field the better you can operate in the field.”
In your experience, what are the key factors that differentiate a mid-level data scientist from a senior-level one in terms of skills and responsibilities?
“The best leaders in data science are those that read about the field as much as they can and as often as they can. A moment ago I emphasized the importance of reading books, but I also recommend blogs and social media. It is also smart and wise to be in the habit to read publications in academic journals. Lastly, I also frequently remind folks that it isn’t necessary to always always always read only the most recent publications. And it is not necessary to read the most advanced publications. Occasionally reading foundational papers and publications from decades ago, which are still valid an applicable decades later, can helpful to professionals looking to grow in the field. And refreshing your knowledge by reading up and brushing up on the rudiments can also be helpful. A favorite book of mine at the cutting edge is Algorithms of Oppression. A favorite book of mine that revisits the rudiments is Burn the Math Class.”
How can a data scientist effectively communicate their value and negotiate for better positions or salaries, especially in a competitive job market?
“As a former attorney I love this topic of negotiations. One of the biggest mistakes in salary negotiation is to think of it as a “money only” proposition. There are dozens of other components to consider.
In my experience it is smart to work through a long list of forms of compensation. The employer might not have flexibility on many points forms. Maybe the employer will not have much room for negotiation on most forms. However, by systematically asking for better terms on each form of compensation you can avoid leaving money and compensation on the table. This process, when done well, can also build and strengthen rapport. The process and discussion need not be contentious.
The items to consider asking about include starting bonuses, retention bonuses, performance bonuses, equity, paid time off, job title, work from home, office supply stipends, commuting costs, parking costs, relocation costs, continuing education, mobile phone stipends, project placement, and even start date. If you want more on these find my article on the topic over at Medium.
Another item to keep in mind on this topic is that the research is clear on how those who negotiate earn more than those that do not negotiate. Also, the process of conducting salary research is low-cost with a very high rate of return.”