Progress Presentation

[westrc]

No description provided.

[nmkwangg]

Interesting ideas. For idea 2, I think that managing and filtering large-scale genomic data requires significant expertise. Ensure you have a clear plan for handling such a large dataset and addressing missing data. And for idea 3, I'm curious about how you’ll evaluate the success of your system. Will it be based on reduced response time or improved customer satisfaction?

[ernbilen]

Very well presented. I think you did a really good job explaining bio topics to non bio people like the rest of us! I like how you talked about each specific gene and what they affected. But maybe they were not as obvious. So, the reason why some that popped up could be obvious is they were studied in the past and determined to have something to do with Crohn's. But imagine if you were the first one doing this study- your results would have been very new. To me, finding patters in our genes like that is already mind-blowing, so you could explore that more. Like, what are some of the things genes can predict vs not. But if you want to focus on Crohn's, then definitely do a lot of literature search because you need to be very familiar with area first in order to conritubute to it later, and unfortunately I'm not... Definitely talk to Bio people too, Roberts, Kusher, Forrester...

[westrc]

Progress Presentation.pptx

[ernbilen]

Great data + setting. It is great that you get to work with real data from the field.

• Prioritizing urgen/non-urgent tickets is a great start. Your model can learn from the past closed tickets and figure out if things could have been done differently and apply that for future tickets.
• Text analysis also almost must-do in your setting. You can consider word clouds, and LDA for topic modeling to try to find patterns in customer complaints.
• You can consider agent level analysis too, analyze which agents are top performers etc. for promotion purposes.