search

stgilhool / IndegoProject

Sample data science project to analyze Indego Bike data
0 stars 1 forks source link

hey chief let's talk why not #6

Open es-wong opened 6 years ago

es-wong commented 6 years ago

yo

noodled on this a bit over the weekend. let's meet up soon to jam--i think it's kinda coming into focus (...at least in my notebook)!

stgilhool commented 6 years ago

added a couple lines to transMatrix.py and pushed to dev. Make sure to pull the changes before you work on it again

stgilhool commented 6 years ago

added some lines (to a renamed copy called tryMatrix.py). It's in steve-dev only. Makes a vector with 1 bike at first station and applies two trips, printing out the probability vector after each. After two trips, there's one station that has like a 6% chance of having the bike, and all others are like 3% or below (with most between 0 and 1%). So there's a clear favorite, even if it's still low percentage. Would be interested in the likelihood vs. distance or population density or whatever else.

stgilhool commented 6 years ago

Pushed tryMatrix.py to dev. I changed it so that it starts with one bike at each station and all bikes make a trip each time. Also settles to a stable state within ~10 trips. Interesting.

es-wong commented 6 years ago

my G my G i got a lot of the visualization stuff working today. i'll edit it up tomorrow and i think it would be prudent if we discuss stuff soon! i guess tomorrow is out due to the impending blizzard, but lemme know when you're around!

bikenet_try1

(this still needs some cleaning up, but i think we're getting there).