My knowledge science journey started with a disaster.
At 9:01am on Monday morning, my telephone buzzed with a PagerDuty alert. Instinctively, I pulled up the corporate dashboard and noticed that mortgage approval charges had doubled in a single day.
My coronary heart began pounding.
I used to be a threat analyst on the time managing a lending portfolio. Approving loans for a higher-risk phase might have a huge impact on loss charge.
Bleary-eyed, I dove into the information.
I frantically:
- Wrote SQL queries to drag knowledge,
- Analyzed knowledge in Python attempting to identify anomalies,
- Appeared for patterns that would clarify the sudden spike in approval charges.
I found that one particular function within the credit score mannequin had drifted.
Seems it was a easy timezone error.
This incident marked a turning level in my profession, and was the start of my journey going from knowledge analyst to knowledge scientist.