A scipy-like implementation of the PERT distribution
A scipy-like implementation of the PERT distribution.
In my current job I work a fair amount with the PERT (also known as Beta-PERT) distribution, but there's currently no implementation of this in scipy. To make up the deficency I crafted up my own PERT distribution class, leveraging numpy and scipy to properly flesh out the functionality. The API is heavily modeled after the scipy.stats methods API's.
TODO: when I figure out how in blazes to add these ;-)
Installation is straightforward:
pip install pertdist
Usage is very similar to what you would find in a scipy.stats class as well:
from pert import PERT import seaborn as sns pert = PERT(10, 190, 200) sns.kdeplot(pert.rvs(10000))
On running this you should see a chart of a heavily low/left skewed distribution (recommended running in Jupyter or Spyder).
- Develop unit tests
- Especially around flexible identification of various data types, eg: accepting DataFrames, Series, lists, etc.
- Build out the following scipy function analogues:
- stats (implemented, but needs refinement)
A version history is located here
Since this is my first published project, I'm pretty relaxed about contributions. Feel free to send me a pull request with any updates/changes/etc you have in mind!
Note that I do follow Vincent Driessen's Git Branching Model rather rigorously. If you do contribute, it'll most likely be pulled into the
Also, I'm rather fond of Semantic Commit Messages, but I'm only picky about those for my own contributions, feel free to use wahtever commit message style you'd like.
This project uses the GNU General Public License.
Short version: Have fun and use it for whatever, just make sure to attribute me for it (-:
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.