The field of Data Science is a young field. So, with no paved road to future success, where are we going? As analysts, we should use facts, analysis and domain knowledge to perform a viable forecast for ourselves, if we are to be serious in forecasting the reality for others! With the history of other historically nascent industries as data, let’s find connections to Data Science. (and hope history will repeat itself!)
I will here connect Data Science with James Bessen’s conclusions (author of “Learning by Doing”[isbn 978-0-300-19566-8]). He argues convincingly about tech fields in general. Using his conclusions, I will try to predict possible scenarios for the young tech field of Data Science.
What will happen to the analytics/Data Science community in a few years?
The 3 observations below are directly connected to Bessen’s observation about where a technology is in its “life-cycle”. Specifically, these observations mean that Data Science is in it’s starting phase (not so surprisingly…). Below that is some educated guesstimates of coming decades.
What are some signs of an immature market?
1: [Education] Analytics are mostly done by “educated professionals”, as opposed to vocationally trained workers, which is a sign of an immature technology. “Educated professionals” adapt constantly to new technology since their education is “learning how to learn”.
2: [Narrow markets] Skills are more-or-less domain specific. It is still very hard to move from the insurance businesses to online retail. Bessen notes how hard it was moving from one type of weavers’ mill to another type, in the first decades of the industrial revolution.
3: [Sharing] Competition is non-existent, as immature markets still expand. Therefore, knowledge sharing and open source has a large community. This non-competition benefits workers AND companies. The early home computer clubs helped Microsoft and Apple founders develop, as an example.
Forecasts of where the market will be in the next decade/s
So, when Data Science/analytics technology matures, where is it headed the next decades? (NB: the word “technology” is both software/hardware and the associated skills)
Forecast 1: [More vocational training] More analysts with vocational training, and specific skills. Today, a teenager can run a big weather simulation on Amazons elastic cloud from her laptop. So, that extra college year won’t be valuable to the employer. Universities adapts to narrow learning-by-doing programs, OR become more theoretical/research oriented. (probably both will happen, and universities/colleges/highschools will complement each-other).
Forecast 2: [Wages might develop] Wages will go up more. Or not. It depends if the analytical problems (and solutions) are transferable between companies. Say you develop a method/model to improve profits for shoe retail stores, but (e.g.) airplane companies won’t pay you for that. But, other shoe retailers will, and at an increased salary. I’m a bit undetermined as to what will happen here.
Forecast 3: [Less sharing] Cooperation online will diminish. After all, with a saturated market, my specific skills are worth more if fewer people know about them. Even to the point that it might be the only way for me to get a job: sharing only means stiffer competition.
Forecast 4: [More patents] Patented technology will be even more of a big deal, as companies do not want to share, either. More non-disclosure agreements and non-compete clauses. (This is seen today, but not as severe, at least here in Sweden).
Forecast 5: [Licensing] Professional licensing will create entry barriers to protect analysts already on the market. The statisticians already have their own system of “chartered statistician”. Why not a “chartered Data Scientist”, or whatever the title will be in the future. Sadly, according to Bessen, a license gives a minimum quality level but does not improve overall service quality.
Forecast 6: [Prices down, salaries up] The overall price of analytics will go down. As the market stops growing, the crunch will create cost-efficiency as a way to make profit. The (vocational) analysts will get a bigger share of profits, as said above, but will also be able to produce so much more. With mature technology, the output and consumption rate is higher as we produce more and buy more at a lower price.
Confounding factors that can prove Bessen wrong
Each of the 3 Data Science competencies (programming/statistics/domain skills) have journeyed from immature to mature technology. Domain skills is a bit wague admittedly, but any technology dependent area go through the same stages. What is interesting/confounding is even if all the 3 competencies have matured individually, Data Science have not started out as a mature application. The field of analytics is more than the sum of it’s parts, and has it’s own development cycle.
General conclusion (please revisit this around 2041)
The mature market for analysts will see consumers and service providers do more analytics at a higher rate. Specially trained roles will streamline the analytical process. More lateral movement of workers between industries, and higher salaries. Competition increases, and entry barriers to the profession will go up.
Now, to be fair, some of the forecasts above is/might be already in progress. I have not seen them radicalised yet, though. So I am, just as you, eagerly anticipating the future, interested to see where we all will end up professionally!
Dan Gustafsson 