sfswift
commented
6 years ago Thanks for the feedback! I think you're right about rephrasing my problem - I've put so much thought into this since starting my job, I think I lose clarity on it since I've had so many different ideas.
We have scrapped linkedin data, although it's proven hard to link to our students (seems we never have up to date emails). We also use careerbridge (careerbridge.wa.gov) for employer data (also limited because you don't have to provide your SSN when applying to the programs).
Surprisingly enough, we don't yet use much data visualization (mostly frequency tables?! ridiculous!), and we've hit quite a bit of push back when trying to implement Tableau. So that alone will probably be a big adjustment, and I'm hesitant to include more right off. It would be great to have sentiment analysis and include more "data science" though! And it works for the purposes of this class.
Again, thank for the feedback! You've given me some good things to think about!
The idea is to link all of the data they have on their students, in order to run analytics which can be used to gain better insights on how they are serving their students.
I think the problem needs to be rephrased. The problem is that you need to gain more insight into student opinions in order to better serve the students. You want to know what you are doing well so you can keep it up, and what you are doing poorly so you can make improvements.
It sounds like you already have all of the data needed for the proposed solution. It is just a matter of linking it all up by student. Is there any other useful data that could be collected (i.e. social media, third party college rankings, glassdoor, linkedin, etc?)
Maybe the proposed solution could bring key positive and negative opinions to the forefront, rather than just providing summaries/visualizations. If you collect more free text data, then you may have to run sentiment analysis, and classifiers to classify strengths and weaknesses.
I like how the idea is based on your real job, and real data that you have. This sounds like a solution that could definitely be implemented, and would have a positive impact on your organization. It is really interesting that this is real. Don’t be afraid to think even bigger than this proposed solution. Think about how you could use more “data science”, even if you don’t know how to execute it yet.