How I Helped Optimize Donut Sales with Amazon Q

My co-worker's homemade donuts are a huge hit at home, especially when fresh from the oven. We would love to start selling them at the office we tried selling them twice and noticed as much as people enjoy them throughout the two days, sales peak when they're freshly baked.

To get the most sales and use the least effort she needed insights & validation on when to bring the donuts & not remain with any unsold ones. Being terrible at sales but a great customer I decided to use my little data analysis skills to see what we can do. To optimize her sales and production schedule, I analyzed the co-working space's check-in logs, examining members' clock-in patterns to determine the best times and days for selling donuts. 

So first thing I did was get clock-in data from the IT department and make sure it was anonymized before doing some analysis.

I took data for the last 3 months: September October and November.

The data we have comprises of Date And Time Personnel ID, First Name, Device Name, Event Point, Verify Type, and Event Description.

From a glimpse, the data is about 8557 logs of every time someone checked in.

The first thing was to anonymize the data to maintain the privacy of the personnel in the office. 

So I asked Amazon Q to help me do this 

I promoted Amazon Q: I need code to anonymize data 

The columns to anonymize are: Personnel ID, First name

 Here is what Amazon Q came up with : 

I then further asked it to rewrite the code to be able to have the anonymized data to a new Excel file feedback message to confirm successful save or report error.

Amazon Q suggested hashlib library to scramble the plain text into unreadable text.

Here is the output and a few modifications to suit my context.

In the first part of the code we utilize salt :
Salting is a security measure that prevents hackers from quickly cracking passwords and data. Before hashing the data, Salt adds a random string of characters to it using os.urandom() .

We then have a function that converts the values to a string and combines it with the randomized output by salt. We then have a function that identifies the columns.
So now I have the anonymized data we can use to make an analysis and find the best time to sell some freshly baked donuts!

let's go!

Here is the output:

It shows that on the top days, we had a lot of traffic on Thursday with 561 logs, and on Sunday have only two logs.

But now we want to the peak morning hour for activity

So I prompted Amazon Q to with the above prompt.

Here is the output it gave :

So Tuesday is definitely the best day for us to for us to get the most sales.