google trends...

36 views
Skip to first unread message

William Stein

unread,
Mar 13, 2016, 3:27:21 AM3/13/16
to sage-ma...@googlegroups.com, int...@sagemath.com
I just spent some time looking at Google trends:

Interest in Sage was steadily declining since 2011... until it finally
ticked up recently around October 2015:

https://www.google.com/trends/explore#q=sagemathcloud%2C%20%2Fm%2F0bcfvp%2C%20cython%2C%20%2Fm%2F0nrfc&cmpt=q&tz=Etc%2FGMT%2B8

The attached image (and link) has google trends for sagemath (the
software not search term), sagemathcloud, and Cython, and Maxima.

- sagemath = software that I hope we are trying to bring back from decline.
- sagemathcloud = new software that is part of my strategy to give sage hope
- cython = spinoff from sage (and Ewing's pyrex) that I launched in
2008, which is a clearly successful project with positive momentum
- maxima = open source software that is coasting down...

As a *reality check* if you put any of Matlab, Mathematica, Maple, (as
compute software) into Google trends, they are at significantly more
popular than everything above. More precisely, Maple, Mathematica,
and Matlab are 5x, 10x, and 100x more popular than Sage.
Interestingly, the relative popularity of Maple, Mathematica, and
Matlab compared to each other, almost *exactly* matches with the
relative number of employees their companies have, which is a good
double check.

Trends? Here's the chart comparing Maple, Mathematica, Sagemath=0 and
numpy. I exclude Matlab since it is so much bigger:

https://www.google.com/trends/explore#q=%2Fm%2F0kl1x%2C%20%2Fm%2F0d2jd%2C%20%2Fm%2F0bcfvp%2C%20numpy&cmpt=q&tz=Etc%2FGMT%2B8

Mathematica, has been declining steadily since 2005... whereas numpy
has been increasing steadily. In fact, numpy and mathematica are
almost neck and neck right now, and numpy will probably pass
Mathematica soon.

-- William
Screenshot 2016-03-12 at 11.09.47 PM.png

William Stein

unread,
Mar 15, 2016, 2:23:36 PM3/15/16
to sage-ma...@googlegroups.com, int...@sagemath.com
Here's a blog post I found today on a similar topic using similar data
with a similar conclusion... but with a focus on data science tools
(still mentioning the Ma's):

http://blog.dominodatalab.com/open-source-winning-against-proprietary-data-science-vendors/
--
William (http://wstein.org)
Reply all
Reply to author
Forward
0 new messages