Start
October 6, 2018 - 5:00 pm
End
October 6, 2018 - 6:00 pm
Address
Online Webiner View mapIDEAS & Data Application Lab co-host this live webinar:
Guest Speakers:
Neal Fultz：
Neal Fultz is an experienced computational statistician at Factual, working with the Audience Insights team, and the owner-operator of NJNM Consulting, a botique analytics company in Los Angeles. He has authored a number of Python, R, and Javascript libraries, and is a core contributor to the DeclareDesign project at UCLA Political Science.
Topics:
Data Science for Good – A Case Study in Wisconsin Higher Ed
or the last six months, we have been volunteering with DataKind (datakind.org) to help colleges in Wisconsin understand why some students didn’t finish their degrees on time, based on historical enrollment data, and what they could do to improve graduation rates overall moving forward.
Collaborating with non-profits and working entirely remotely had different challenges compared to our normal 9-to-5 jobs. We will discusss the data and challenges, both technical and operational. On the technical side, there were many challenges related to data quality, overfitting on noise, and algorithmic bias. On the operational side, there were challeges handling PII securely, preparing for restated data, presenting results to not-necessarily-technical-stakeholders, and providing useful data science artifacts that can be referenced, reused, and provide lasting value. Thanks to support from DataKind, we were able to work our way through these issues and complete the project successfully and on time.
We will provide suggestions based on those experiences so that non-profits (and other clients) can be better prepared for data science consultants, and so that data scientists can get involved and improve their impact when they volunteer in the future.
About Data Application Lab:
About IDEAS;
