search

cscareers-dev / suggestions

0 stars 1 forks source link

#2022_grad_process #9

Open tristingreenstein opened 3 years ago

tristingreenstein commented 3 years ago

What is your suggestion?

Do not allow !process company reject|phone|oa|final|offer without having a !process company apply from the same user. It skews the stats on the website

What is the value add for adding this to the Discord server?

Fixes the skewing of stats on the website

Additional context

No response

rieryn commented 3 years ago

it would be better for !process to backfill previous steps, if you have per user tracking

joeydotdev commented 3 years ago

it would be better for !process to backfill previous steps, if you have per user tracking

When implementing this functionality, I was debating whether or not to backfill previous empty steps if they did a final step. What I realized with this is that not all companies have the same processes (not all companies have OAs - so backfilling an OA event doesn't make sense).

But in regards to @tristingreenstein 's suggestion, I think the lowest common step when someone does receive a higher level step (phone, final, offer) would be an apply step - so I think backfilling applies make sense.

When I have some free time, I can run a query to get the number of people that:

Based on how large this number is, we can get an idea of how beneficial this is. Current architecture sort of makes this a little more involved ask. Will get back to this.

tristingreenstein commented 3 years ago

just as an additional comment, any company than has more offers, rejected, phone, oa, or finals than applies is automatically listed as skewed data. Not just offer vs apply Microsoft 38 apply 52 offer 47 rejected - min 9 person skew (difference between apply and rejected) Amazon 51 apply 88 offer 80 rejected - min 29 person skew VMWare 12 apply 27 offer 18 rejected - min 6 person skew Imc 21 apply 48 rejected - min 27 person skew 15 of the first 30 companies have skewed data