A simple library for estimating the impact of an intervention, with humility
Estimate the impact of an intervention, with simplicity and humility
You can install with pip install preposterous
If you want to install from source, then clone this repository and run python setup.py install from the project root.
Pre/Posterous is under active development and the current release can be considered a ‘proof of concept’. The largest restriction is that it only imports data from the Reporter app, which is my current default recommendation for anyone who is trying to track something.
Quantified Self data
The primary use case is for quantified self, where you have periodic measurements of the target metric (weight, categorical sleep quality, ect) and potential interventions (medications, diet shifts, ect). This library can organize these into ‘natural experiments’ that point the way towards a causal relationship
Warning Python is pretty great, but nothing can replace a well powered Double Blind Randomized Controlled Trial for establishing causality. That said, many (most?) situations do not lend themselves to RCTs, and yet we’re still forced to make decisions. That’s where tools like this, used with an appreciation for non-binary modes of belief, can be helpful.
import preposterous.preposterous as ppl pdf = ppl.PrePostDF() pdf.add_outcome(filename='data/sample_reporter_output.csv') pdf.add_intervention(filename='data/sample_reporter_output.csv') print(pdf.basic_info()) pdf.generate_confusion_matrix(intervention='Exercise')
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
|Filename, size||File type||Python version||Upload date||Hashes|
|Filename, size preposterous-0.0.1-py3-none-any.whl (4.2 kB)||File type Wheel||Python version py3||Upload date||Hashes View|
|Filename, size preposterous-0.0.1.tar.gz (4.0 kB)||File type Source||Python version None||Upload date||Hashes View|
Hashes for preposterous-0.0.1-py3-none-any.whl