A few websites catering to analytics and data science professionals have experienced tremendous growth recently. Organizations such as INFORMS or AMSTAT have seen their traffic explode, targeting high school students to join the ranks of data scientists. Niche publishers providing high quality, actionable content - and run by true data scientists rather than journalists - have also seen spectacular growth.
Figure 1: Number of monthly visitors for one of our channels
By data science, I mean all disciplines focused on optimizing value through data analysis. It includes operations research, machine learning, data engineering, biostatistics, data mining, business analytics, predictive modeling, data plumbing, statistics and many more. For a full list, check the following articles
The chart in figure 1 shows growth for one of our channels (representing a fraction of all our web traffic), after filtering out non-human traffic. While in our case, growth is partly fueled via growth hacking techniques, much of it is still organic. In figure 1, you can see periodic dips in traffic (December and July), a spike around March/April 2014 when my book was published, and a spike in March/April 2015 (well, that one is not visible yet) when we will release data science 2.0..
As this point, the fake data scientist bubble has or is in the process of exploding - so we are losing this traffic, which is great. The traffic that we lost is more than compensated by new traffic generated thanks to new trends, including
Indeed, we even thought of having members pay a tiny symbolic fee, not to generate revenue, but to filter out members that do not bring value to our community. We won't do it, instead we focus our efforts in attracting high quality members. All our ads target US-only professionals, including those aimed at recruiting new members. And we are in the process of launching a few new channels:
We are also careful about not inflating our traffic statistics as reported by Alexa. In particular, we do not split our articles into multiple pages to boost page views count, as this would be unfair to advertisers purchasing display ads. So our numbers are conservative.
Figure 2: Growth for our main LinkedIn group, over several years (now 150,000 members)
Figure 3: Twitter growth on our main account
Other indicators of growth include keyword trends for data science related keywords (number of searches on Google), and number of job openings (there's data available on Indeed, but it looks wrong). Finally, figure 1 offers the most realistic picture: growth rate is constant. LinkedIn and Twitter charts (showing increasing growth rate over time) do not take into account inactive users.