1. Increased need for advanced statistical tools beyond just Excel and more affordable (and more modern) than SAS. This means that R is getting more and more attention from business analytics folks, just as Python is as well. If someone lives only in one of these communities, it's possible to misinterpret this overall expansion as being specific to just their language/ecosystem.
2. More heterogeneous data streams and increased data volumes mean that traditional the data warehouse pipeline model of ingest -> clean -> structuring/modelling -> analysis is breaking down. (Cloudera, for instance, acknowledges their evolving customer needs in the DW space by branding their Hadoop stack as being an "Enterprise Data Hub".) However, most people are not sold on the idea that there is a Single Unicorn Vendor that will sell them a single silo or comprehensive BI suite that fixes everything. Instead, in the face of increased demands for agility and increased heterogeneity, ad-hoc data munging is a real need. This is an area where Python truly excels.
3. The author of this blog post is squarely focused in the world of stats. This is R's home turf and of course, R and S have been the de-facto languages of academic statisticians for decades. By the time that a user like him would consider turning to Python, then R's days will truly be numbered. That has not happened yet, nor is it the contention of the original blog post he was responding to. Instead, it's the idea that there is a huge diversity of statistical and analytical needs in the world, and Python is rapidly being adopted in those spaces because it's easy to pick up and it is a very polyglot tool.
The author of NumPy has commented that, "R is still king in the statistics department (and some social sciences). Everywhere else in the university - engineering and sciences - people use Matlab, C++, and more and more, they are turning to Python."
1. Increased need for advanced statistical tools beyond just Excel and more affordable (and more modern) than SAS. This means that R is getting more and more attention from business analytics folks, just as Python is as well. If someone lives only in one of these communities, it's possible to misinterpret this overall expansion as being specific to just their language/ecosystem.
2. More heterogeneous data streams and increased data volumes mean that traditional the data warehouse pipeline model of ingest -> clean -> structuring/modelling -> analysis is breaking down. (Cloudera, for instance, acknowledges their evolving customer needs in the DW space by branding their Hadoop stack as being an "Enterprise Data Hub".) However, most people are not sold on the idea that there is a Single Unicorn Vendor that will sell them a single silo or comprehensive BI suite that fixes everything. Instead, in the face of increased demands for agility and increased heterogeneity, ad-hoc data munging is a real need. This is an area where Python truly excels.
3. The author of this blog post is squarely focused in the world of stats. This is R's home turf and of course, R and S have been the de-facto languages of academic statisticians for decades. By the time that a user like him would consider turning to Python, then R's days will truly be numbered. That has not happened yet, nor is it the contention of the original blog post he was responding to. Instead, it's the idea that there is a huge diversity of statistical and analytical needs in the world, and Python is rapidly being adopted in those spaces because it's easy to pick up and it is a very polyglot tool.
The author of NumPy has commented that, "R is still king in the statistics department (and some social sciences). Everywhere else in the university - engineering and sciences - people use Matlab, C++, and more and more, they are turning to Python."