Data Science · Analytics
So You Want to Be a Data Analyst?
By David Robertson | February 07, 2019
It has taken me 20-plus years — five years in for-profit and 16 years in non-profit — to feel comfortable and confident with the title of “analyst.”
I am a self-professed data analyst with 20-plus years of experience in the field of decision science and prospect research analytics. During my time at Syracuse University as director of operations research, I held the position as adjunct professor as well as at Le Moyne College, teaching statistics, applied data mining and data science. Currently, my title is decision scientist at Upstate Medical University in Syracuse. I am finishing what I call the “capstone” phase of my amazing career.
Yes, it’s taken me more than 20 years, but I can say with all the confidence in the world that I am an analyst — a data analyst, and proud of it.
I am passionate for the profession and serious about the path taken to obtain the title of “data analyst.” What an amazing feeling it is to apply true methods and a calculated approach to separate what Nate Silver calls the “signal from the noise.” To have the knowledge, creativity and confidence to help management make better decisions, and development officers find that golden needle in the haystack.
The title of “analyst” is becoming more and more popular. With the exponential growth of data collection and the desire to mine and understand known trends within the data, we need more analysts. However, in my opinion, the title of “analyst” has been used, reused, misused and watered down to the point where it has lost its meaning. It has taken me 20-plus years — five years in for-profit and 16 years in non-profit — to feel comfortable and confident with the title of “analyst.”
And so, for this reason, I’ve put together what I believe to be a solid 10-step path for the aspiring data analyst.
10 Steps to Become a Data Analyst
- Get educated: Everything starts with education. For the aspiring data analyst, a four-year degree with a major or (at a minimum) a minor in data science is highly recommended. For the liberal arts and social science majors, a four-year degree with exposure to data science would introduce you to descriptive statistics, forecasting and classification models — the basics for prospect analytics. You would have a highly desirable qualitative and quantitative academic background.
- Get even more educated: I believe it’s necessary to obtain a graduate degree. It is possible to land an entry-level data analyst position with only an undergraduate degree, but a graduate degree is a must for a more sustainable career path. Any employer or future employer worth their weight would require a graduate degree, and may even help with tuition costs.
An MBA with a concentration in applied statistics would set you apart. Applied statistics offers real-life scenarios and applies concepts and methods to problem solving. These demonstrated methods (not theory) apply to all industries. As a professor, I always tell my students, “I don’t care what industry you’re in; if you’re making penicillin or space shuttle wings, methods are methods. The analytical approach and methods used are the same. We’re just asking and answering different questions.”
- Build your creativity: I have found that the best analysts in any industry are those who are creative in their approach. They approach the chaos of data as an artist might approach a canvas. They ignore the noise and acknowledge its possibilities — its potential signals.
As competent and creative analysts, we know what methods to use to bring about stories we anticipated or those that are newly discovered.
This creative approach comes about by having intellectual freedom and confidence in your own abilities, as well as confidence in the data. Having solid background in analytical methods, confidence in the methods you employ and the freedom to explore new analytical concepts give you confidence and a high level of passion for the work you do. Passion (Y) = Education(x1) + Confidence(x2) + Creativity(x3).
- Discover and retain passion: To be content and successful as an analyst, you must discover and retain your passion for the profession, at least if you’re going to stay in the field for the long haul. Passion comes about with project success, recognition of good work and self-satisfaction. Passion for your work is derived through self-awareness and self-study and is self-driven. No one gives you passion; it finds you through your dedication and hard work.
- Find mentors: A mentor is an amazing luxury to have. I was lucky early in my career to have teamed up with a true analyst and mentor. This gentleman, David G., is a brilliant, eccentric individual and willing to share his discoveries. There was no holding back on his end — if I wanted to learn, he was there to teach me.
David showed me the power of MS Excel: its formulation and functions. Having a solid understanding of MS Excel and its powerful analytical functions, along with its unending potential, helped in labeling me as the go-to guy for questions and modeling-building. As an adjunct professor, I too shared what he taught to me. I brought his methods and approach into the workplace and the classroom…those things you can’t get from a book or manual. Unfortunately, when my time came to an end with this employer, so too did my relationship with David G. The good news was, in the non-profit prospect research world, we have an individual who became a big advocate of mine, if not mentor: Peter Wylie.
Peter is a well-known practitioner in the fundraising analytics world, and rightly so. His contributions to our profession are many. He is a highly sought-after writer and presenter at many conferences, both domestic and international. Some consider Peter the father of fundraising analytics. I quickly reached out to Peter and we became colleagues and friends. His positive and passionate approach to fundraising statistics was addicting. He was always available to kick around an idea and always willing to correct me when I misspoke during one of my many conference presentations. His observations, insights and knowledge are a goldmine and always welcome.
True mentors are selfless individuals who want nothing more than to share their knowledge with you. There is no competitive nature in their approach.
Unfortunately, mentors can be hard to find. So what do you do when you don’t have a mentor? Find one through literature or online. Four individuals I never get tired of listening to or read about are: W. Edwards Deming (Quality), Nate Silver (Statistician/Politics/Sports) and his FiveThirtyEight website, James Harris Simons (Mathematician/Code Breaker), and David Blackwell (Statistician/Game Theory). Check them out — they’ll inspire you!
- Know your data: At Syracuse University, I was very happy with the data that was available. Compared to other institutions, the data was clean and well organized. It was not perfect data, but overall it could be used to paint a picture of population trends and giving behaviors. Over the course of my 15 years at Syracuse University, I retrieved and analyzed hundreds of thousands of rows of data. I knew the data and its locations well, and I was confident in it.
The lesson here is to familiarize yourself with how your data is organized. Get to know the data architecture and get to know your data warehouse people.
- Get out of your cube and promote: The days of data analysts being reclusive or confined to a cubicle eight hours a day are coming to an end. We have been our own worst enemies when it comes to promoting our work. We have to take charge, ask to be part of managers' meetings, and highlight our past success and future possibilities.
Beware, however: The language of the data analyst/statistician can be like fingernails on a chalkboard to others. Managers do not want to hear the statistical methods used (p-values, t-test, regression, decision trees and classification). Keep it simple (a “green is good/red is bad” approach) and promote yourself and your project success every chance you get.
- Be proactive: Reach out to colleagues with whom you haven’t worked but would like to. Provide them with a report that others have requested from you — a report that others have found helpful.
I have found that seasoned development officers are far less likely to reach out compared to newer development officers. In many cases, the seasoned development officer is happy with the methods they’ve used for years. The newer development officer may be more current on machine learning techniques and comfortable learning a new approach. Regardless of tenure, colleagues will always be grateful in any help you can offer.
If you find a lag in your day, create unique databases and explore new methods to examine data. Watch videos on YouTube and educate yourself on advanced functions offered within MS Excel. Listen to lectures on game theory, forecasting, classification, text-mining and decision tree analysis. Discover the random number generator within the MS Analysis ToolPak in Excel and create your own database to apply your learned methods. (Feel free to reach out if you’re unaware of MS Analysis ToolPak).
- Join, speak and share at conferences: Attend conferences and share your knowledge. If you’re uneasy presenting solo, offer to co-present. Become confident in your abilities — you are the expert, and the audience is hungry for your insight and knowledge.
My greatest memories are of speaking at local, national and international Apra conferences. Your colleagues are eager to learn and you have more to offer than you think. Colleagues will become energized by finding their approach has been validated. You will instruct, inspire and, much like teaching in a classroom, learn along the way, as well. Share your knowledge!
- Stay current: One thing I’ve learned along the way: The field of analytics is growing exponentially. However, keep this is mind: Stats is stats is stats. A solid background in statistics will take you anywhere you wish to go. Pick a discipline and a solid statistics background will get your there.
The systems and tools we use are outpacing our ability to keep up. Become knowledgeable and comfortable with a few of the statistical software packages and stick with it. I’m old-school; I use MS Excel and SPSS primarily. I know a little R to get into trouble. Don’t laugh: Nate Silver predicted the Obama election in 2008 with only the use of MS Excel. Good enough for him, good enough for me. It all comes down to a basic understanding of applied statistics and proper methodology. And never forget to always love what you do!
In Conclusion
Every day, I utilized statistical methods to deliver true data science to guess work. Through my analysis, I tried to extend power and confidence to the development officer. There was personal satisfaction in that confirmation; applying methods to subjectivity. Many times, data mining would bring to the surface that golden needle, that signal, that was hidden in the thick haystack of noise.
My time at Syracuse University was one of learning, professional growth and exploration. This 10-step road map is a result of those 15 years. At SU, I was given the opportunity to use my education and creative thinking to shed quantitative insight and direction to our fundraising team. The job title of “data analyst” is one that is rich with assumptions (and rightfully so). Assumptions that can be supported include; having the knowledge to employ correct scientific methods to the question that is being asked (descriptive and/or predictive).
The true “data analyst” has the knowledge as well as the confidence to explain the rationale behind the methods utilized and the conclusions extended to the end user.
David Robertson
Decision Scientist, Upstate Medical University
David E. Robertson, Jr., holds the title of Decision Scientist at Upstate Medical University, where he works with leadership across the campus. He helps to identify strategic priorities and growth opportunities, develop business plans, conduct market analysis, and monitor portfolio metrics.
David was the former director of operations research at Syracuse University for Prospect Identification/Analytics. In that role, he employed a variety of descriptive, predictive and prescriptive analytical techniques and data visualization to provide information, analysis, tools and insight to support management decision-making and effectively communicate data and insights to stakeholders.