Yannick Pouliot
14 followers
14 followers
Yannick Pouliot's posts
Post has shared content
Very good advice. In my experience, the vast majority of academic scientific code fails miserably with respect to ...
Write programs for people, not computers.
Use the computer to record history.
Use version control.
Document design and purpose, not mechanics.
As for "Collaborate", I don't know of an instance or pair programming ...
Write programs for people, not computers.
Use the computer to record history.
Use version control.
Document design and purpose, not mechanics.
As for "Collaborate", I don't know of an instance or pair programming ...
Best Practices for Scientific Computing
For those, like me, that weren't born developers, here there are 10 principles to be better at scientific programming.
"Write programs for people, not computers"
http://andreacirilloblog.wordpress.com/2014/11/05/best-pratice-for-scientific-computing/
For those, like me, that weren't born developers, here there are 10 principles to be better at scientific programming.
"Write programs for people, not computers"
http://andreacirilloblog.wordpress.com/2014/11/05/best-pratice-for-scientific-computing/
Post has attachment
A nice way to navigate R packages by broad area of function: http://mran.revolutionanalytics.com/taskview/info/?MetaAnalysis...
Post has attachment
investigating t-SNE algorithm. Interesting...
http://en.wikipedia.org/wiki/T-distributed_stochastic_neighbor_embedding
http://en.wikipedia.org/wiki/T-distributed_stochastic_neighbor_embedding
Post has attachment
Important work
Post has attachment
Nerd friends, might have you used/seen good biocomp applications of the t-SNE algorithm? This is a relatively new algorithm that seems to have decent applicability for non-parametric, non-linear dimensionality reduction...
http://en.wikipedia.org/wiki/T-distributed_stochastic_neighbor_embedding
http://en.wikipedia.org/wiki/T-distributed_stochastic_neighbor_embedding
Post has attachment
A very nice analysis of a nasty problem, namely, how to encode laboratory protocols in a way that promotes experimental replication and metadata exchange by transmogrifying protocols into computable objects.
Post has shared content
Hilarious!
Enjoy the day
beeing different isn't a bad thing. different is just different.
http://goo.gl/iEnpIh
beeing different isn't a bad thing. different is just different.
http://goo.gl/iEnpIh
Post has attachment
Geek friends, anybody using/looking at SciDB, the latest from the fertile mind of Michael Stonebraker?
http://escience.washington.edu/get-help-now/introduction-scidb
http://escience.washington.edu/get-help-now/introduction-scidb
Post has attachment
Good stuff!
Post has attachment
Interesting work that address the nasty problem of uniquely identifying data, e.g., from a publication to the dataset as deposited in a repository.
Wait while more posts are being loaded