Placeholder for ideas for improvement

possibly-sam / mcfc01

Monte Carlo's Flying Circus

MIT License

1 stars 5 forks source link

Placeholder for ideas for improvement #1

Open possibly-sam opened 5 years ago

possibly-sam commented 5 years ago

Better filtering
Static calculation of confidence interval based on assumption of discrete distribution.
UI improvements
Integration with other tools, e.g. MS Power BI

raminty commented 5 years ago

Static calculations are possible if we assume independent risks, and could be interesting to look at. Also:

Highlight to the user (in UI) possible typo mistakes and other outliers in the risk register, along the lines we found while cleaning data
Experiment with ways of getting users to identify if risks are correlated (dependent), especially useful for making the most of MC simulation

possibly-sam commented 5 years ago

Good point. I am thinking that for #1 it is almost a separate application. That is, the application has two almost independent parts. 1) Input and clean the data, and 2) Perform calculations and report on clean data. From that point of view, the typo question might fit in #1.

For #2, i was thinking that if we had detailed descriptions of the risks, we could do a textual analysis for similar terms and assume some correlation. If two risks both have a phrase like 'LTL Trucking Delivery' in them, it might be reasonable to assume a correlation.

raminty commented 5 years ago

Yes, both of those make sense to me. Those textual indicators can also be used to indicate to users if they are possibly duplicating risks that already exist - human-in-the-loop augmentation rather than full automation.

possibly-sam commented 5 years ago

Meeting Room Booked

Holborn - Meeting room 410 St.Giles, 154 Southampton Row, Holborn, London, WC1B 5AL Date: Saturday 11 May Start time: 11:00AM Duration: 4 Hours Guests: 4

possibly-sam commented 5 years ago

Faster(?) version updated to https://north-loop-designs.shinyapps.io/mcfc03/

possibly-sam commented 5 years ago

Initial thoughts on the Black Swan Lever..

bsl-01.pdf

aangt321 commented 5 years ago

Guys where is the link to the clean set of data for download & analysis?

possibly-sam commented 5 years ago

https://github.com/possibly-sam/mcfc01/blob/master/risk_KA_11.csv

risk_KA_11.csv was the file that Ken prepared and that we used in the last hackathon. It is the one I have continued to use for all my testing.

risk_KA_11.zip