Seeking to Bounce from Analyst to Scientist
Hey there,
I’m presently a knowledge analyst with expertise based in SQL and visualization instruments (Tableau & AWS QuickSight). I presently have a task that enables me an excellent little bit of flexibility and free time and I wish to use that to be productive and proceed my profession development. I’m contemplating graduate applications just like the Syracuse for knowledge science.
Some background, I actually haven’t got the maths expertise to be a knowledge scientist. The very best stage of math I took in faculty was statistics. Nonetheless, I do have 8 years of expertise creating KPIs and dealing on the evaluation group at a number of begin ups.
Any recommendation is appreciated.
Comments ( 4 )
The main difference between a data analyst and data scientist is your job scope, then only your skills.
A DA analyse historical data to extract insights. They need to be good at descriptive analytics. They answer business questions based on what happened in the past.
A DS builds model to predict outcomes that will bring positive impacts to the business. Be it reduce churning of subscribers, reduce staff resignation, or increase the CLTV.
You don’t need a graduate program (or Masters / PhD) to become a data scientist. You may read my other posts here if you want a more detailed explanation.
Since you are experienced at defining metrics and leading a team, you should be fine.
Also, I didn’t really need a lot of maths and stats throughout my career (data miner → research analyst → research developer → lead developer → Startup CTO → ML Engineer → Chief Data Scientist).
Most of the time I am required to pick the right model, build it, and deliver business results.
Want to transition faster? Ask your current job to please change your title from data analyst to data scientist (without changing your actual tasks!). That it is really important for you as that is what you are studying, and if questioned that you dont expect them to pay more and/or that you expect to apply what you learn on the job when possible particularly in applying more statistical rigor to your analyses. If they like you and there isnt a lot of red tape, they should consider it a minor change that will make you happy and cost them nothing.
There are plenty of DS roles that lean more on the analytics side. More importantly, the title seems like such a big mountain to reach that needs more education but it’s not. Once you get the title, you can focus on studying for the type of role you want with less pressure but meanwhile you are on record getting experience as a DS.
The line between data analyst and data scientist is in reality pretty blurry. I’ve seen data analysts do pretty complex stats work and modeling and data scientist just do dashboards and simple queries. Also, the DS role is not standard. For example, a lot of data scientists dont do any models (ex – product data scientists focusing on experimentation).
Disclaimer – I’m formally trained as a mathematician, but have taught myself a fair bit of predictive modeling.
I may be biased, but I think to be considered a data scientist, you ought to understand the mathematics and statistics that the models utilize, as well as the technical chops to understand their implementation in at least one programming language.
That is not to say this is necessary in industry. If you just want the title, some other answers shed light on that. But it is only a job title and doesn’t reflect the fact that you’d really still be a data viz technician/analyst and not a scientist (and there is nothing wrong with being an analyst).
However, the graduate program is likely to build that understanding (maybe superficially?) and bring you to a new level of understanding. I’d probably recommend learning/reviewing calculus, probability, and linear algebra, as well as some programming language (python, R, scala…), as I can only imagine these programs would assume familiarity with those fundamental subjects and build from there.
All that said, I think furthering your education is a great idea. Learning is never a bad decision!
It really depends on the domain knowledge area you want to work in as well. Im a research data scientist at Meta. The position itself requires a PhD in a quantitative field (stats, physics, ect) with proven publications in ML related work. We spend a lot of time theorizing new models and implementing/testing them with scientific rigor rather than using packaged models to get insights. So even within the DS field, there can be a lot of variation. Math is a lot more important in the work I do, but generally, if you have basic stats knowledge and are willing to learn more to understand your model outputs then it should be no problem transitioning to that role. Obviously, you should be proficient in Python /R or both as well.