Tagged: Predictive Analytics
March 13, 2019 at 10:24 pm #4353
After some internet surfing, I found this article to be very applicable to what we’re studying in ISyE412 at the moment…it’s related to UW-Madison as well! Might expand on this for our course project if I can find the datasets.March 21, 2019 at 5:48 pm #4367
Thanks for sharing. This is very interesting to read and glad to see other UW schools can benefit from data analysis.March 26, 2019 at 1:31 pm #4377
“The UW System Board of Regents signed in December a $10.8 million contract with Education Advisory Board, buying themselves five years of access to the company’s predictive analytics software, which spits out a student’s “risk score” based on demographics, test scores and high school transcripts. Faculty members and advisers use the scores to reach out to students before they withdraw from the university.”
Wow, this is interesting. I wonder if there is a correlation between preparation received based on demographics, and prior performance as indicators of future performance. For example, if one is from a low-income neighborhood and didn’t get to prepare too well for college, but did well on SATs, graduated valedictorian of high school, and gained admissions to the university, will this student be successful or not based on the presented information?
Excellence on exams and scores in high school and intro level courses might not transfer to overall college performance. This is interesting! Wonder what other applications might be.April 6, 2019 at 1:24 pm #4391
“UW System President Ray Cross said the company’s software is “no magic bullet,” but other universities have seen success, particularly in graduation and retention rates of students who are first-generation, underrepresented minorities or Pell grant recipients. Those groups have historically graduated or re-enrolled at lower levels than others.”
This paragraph highlights crucial methods of how segment data and properly analyze it to attain relevant results. They used classifications and tracking patterns to properly identify students that might need more assistance in staying in school.
I wonder if they can further segment each data into what dorms have the highest dropout rates. Additionally, they could divide the dropout rates by degree type. I think further sample size segmentation could help the school better understand what students are in need of the most help.April 7, 2019 at 11:21 pm #4399
This is pretty cool!!!April 13, 2019 at 12:53 pm #4432
Thanks for sharing this article! I think it’s a really interesting look at big data that we haven’t previously talked about in class. The real-world applications we’ve discussed seem to mostly trend towards sales and how to market towards certain groups of people. It’s refreshing to hear about an application that helps at-risk students, as well as treating the data as if it were protected by HIPAA.
You must be logged in to reply to this topic.