Monte Carlo Scheduler

Project to use Monte Carlo Simulations from historical data to estimate project completion dates.

#Usage Python 3 is recommended! There's some upredictable bugginess with Python 2.7 that I haven't had a chance to debug yet. Use at your own risk. (Though, I guess that also applies to Python3 :)) To run the model based on historical and future data, you have two options. ##To load files containing historical or future data: Use the runModelFromFiles() method as follows:

results = runModelFromFiles(historicalFileName,futureFileName[,verbose=False,trials=10000,plot=True])

The verbose, trials, and plot parameters are optional and have sensible defaults (not verbose, and 10000 trials, True - draw plot).

##To supply your own data as lists: Use the runModelFromData() method as follows:

results = runModelFromData(historicalData,FutureData[verbose=False,trials=10000,plot=True])

Same optional parameters apply from above. Input data must be in format:

historical: [[name,est,actual],[name,est,actual],...]
future: 	[[name,est],[name,est],...]

##Output

Running runModelFromFiles will yield something like this:

>>> MCSched.runModelFromFiles("historical.txt","future.txt")
Running in quiet mode. For full output, supply verbose=True as a parameter. (Warning: it's noisy!)
Running 10000 trials. Set parameter trials to customize, default is 10000
Estimated Total: 48
Min:38.40 (80.00% of estimated)
Max:138.00 (287.50% of estimated)

If plot=True, it will also display a PyPlot graph like this one: The vertical dotted line indicates the sum of the original estimated hours

If you don't assign the return value to a variable it will barf up the entire list of predictions.

#Input File Formats CSV or TAB delimited should work, but I haven't tested extensively. Header row is optional.

"Task 1",8,10
"Task 2",8,6
"Task 3",15,17

The file for future work doesn't require the third column, obviously. For more info, check out sample files testData_historical.csv and testData_future.csv.

#Function Breakdown No promises this is up-to-date. Abandon all hope, ye who enter here.

##loadData(historicalFileName, futureFileName) Loads historical and future data from a file. Input: historicalFileName = filename of file containing historical task data futureFileName = filename of file containing future task data Output: historical = list of historical data in the format: [[TaskName, estimated, actual], ...] future = list of future data in the format: [[TaskName, estimated], ...]

runSimulations(historical,future,n=1,verbose=False):

Runs n simulations of future data based on historical data.
Input: 	historical 	= list of lists in the format: [[TaskName, estimated, actual], ...]
		future 		= list of lists in the format: [[TaskName, estimated], ...]
		n 			= number of simulations to run
Output: 

runSimulation(historical,future):

Runs a single simulation of future data based on historical data.
Input: 	historical 		= list of lists in the format: [[TaskName, estimated, actual], ...]
		future 			= list of lists in the format: [[TaskName, estimated], ...]
Output: predictedTotal 	= total cost of simulated tasks 

summarize(data,verbose=False):

Summarizes a list of values with duplicates to a list of values and their counts.
Input: List of predictions
Output: List of lists; Predictions and their counts: [[prediction, count], [prediction, count]...]

computeConfidence(data,verbose=False):

Comp
Input: Prediction counts, a list of lists in format: [[prediction, count], [prediction,count]...]
Output: Predictions with confidence percentages in format: [[prediction, percent], [prediction, percent]...]

runSampleModel():

Runs a model for the sample CSV files provided.

runModelFromFiles(historicalFileName,futureFileName,verbose=False,trials=10000,plot=True):

Runs a model based on historical and future data read from files.

runModelFromData(historical,future,verbose=False,trials=10000,plot=True,plot=True):

Runs a model based on historical and future data provided as lists.