sweet hatch
Hey all,
I'm here to ask a computational problem (ie I plan to solve it algoritmically) but I believe it has a stats modelling / probability aspect to it and I want to explore it but I am too much of a noob.
I would appreciate some help in exploring this.
Here is the problem statement:
We sample n students for the date they passed a test, how long it took them to prepare for the test, and how many lessons they took during that time.
This is done over the entire year.
I want to know the answer to the following questions:
a) how lessons were being taken in each calendar month of this year
b) how many unique learners were taking lessons in each calender month of this year
this is to map out a hotspot chart of which months were the most.. studious(?)
Computationally, I plan to solve this by imposing some upper bounds on how many lessons can be taken in a day in total and how many lessons each student can take in a day.
My data looks like this FWIW:
<student n><lesson count l><weeks elapsed w><date of test (DD-MM-YYYY) t>
(student1, 10, 1, 3/11/2025)
(student2, 20, 5, 1/11/2025)
(student3, 12, 12, 20/10/2025)
Additional questions:
- The actual data reports time elapsed in units including days, weeks, months. I have rounded them all to weeks. Do I use error bars (or any other method?) to represent this uncertainty? How?
Attached below is a simplified model I've thought about. My thought process here is the student could have taken the lessons any time in the period between their first lesson (~w weeks back from the completion date)
Unfortunately I don't really know where to go from here. If anyone gets what I'm going for I'd apprecicate any insights.
Let me know if I'm completely off the mark trying to model this as a probability distribution. I'm thinking something along the lines of "whats the probability this day was the busiest?" but then again I have no idea what I'm doing exactly so I'm willing to hear out anyone with experience.
Thank you all,
Kamala
latent stone