[School-of-data] School-of-data Digest, Vol 12, Issue 17

Thu Mar 14 15:54:09 UTC 2013

Hi Russ,

The structure of the assignment can also help students be more
productive.  For example, have a few intermediary deliverables that
build towards the final assignment and require that students change
defined roles for each stage (a different leader for each
deliverable).  If no one has made progress perhaps they simply don't
know how to break down the assignment into manageable chunks.

Guides for students about working in groups:
http://writing.colostate.edu/guides/guide.cfm?guideid=42
http://www.speaking.pitt.edu/student/groups/smallgrouptips.html

I find the scientific method to be a helpful framework for data
analysis.  It reinforces the concepts of scientific methods but it
also teaches people to think strategically and critically: Formulate a
question, develop a research design (lit review, hypothesis
generation, set up evaluation parameters), experiment (measure,
collect / find data, process data), analyze and interpret.
Alternatively, a mystery novel framework might be enjoyable and
encourage them to think creatively and independently.

While the subject matter of some of these may not be of great interest
to students in a computer science class, I like these examples of data
use:
- CM Kumar's Masters thesis in City Planning at MIT is about GIS
Methods for Screening Potential Environmental Justice Areas in New
England http://web.mit.edu/dusp/uis/www/projects/theses/chitrak_final_s2002.pdf
- Urban Institute's Catalog of Administrative Data Sources for
Neighborhood Indicators http://www.urban.org/publications/411605.html
- MS Kearney and PB Levine use multiple data sources with an economic
lens to look at teen birth rates (Why is the Teen Birth Rate in the
United States So High and Why Does It Matter?)
http://www.econ.umd.edu/research/papers/608
- S Stephens-Davidowitz's THE EFFECTS OF RACIAL ANIMUS ON A BLACK
PRESIDENTIAL CANDIDATE: USING GOOGLE SEARCH DATA TO FIND WHAT SURVEYS
MISS. http://www.people.fas.harvard.edu/~sstephen/papers/RacialAnimusAndVotingSethStephensDavidowitz.pdf

Hope it all works out for what is left of the semester!

-Johanna

> On 3 March 2013 18:54, Russ Abbott <russ.abbott at gmail.com> wrote:
>
>> A colleague, who is teaching a data science course was called away by a
>> family emergency. I was asked to substitute for a couple of weeks. The
>> students were assigned to do a term project. I spent the first of my two
>> weeks having them discuss their status. It was disappointing. Many hadn't
>> started; any of the rest hadn't gotten very far.
>>
>> I'm looking for some good case studies that I can present to show them how
>> to do a good data science project. Ideally the case study should describe
>> the original data, the tools used, and the steps taken in enough detail
>> that someone could reproduce the work.
>>
>> The course is for computer science students who have virtually no data
>> science background. So ideally also, the case study should not presuppose
>> any sophisticated statistical competence.  Most of them are using Weka.
>>
>> Any suggestions would be appreciated.
>>
>> Now that I write this it strikes me that this is probably a good way to
>> teach a data science course to CS majors.  Work though a number of case
>> studies that illustrate the important techniques and explain how the
>> techniques and theories work as they are used.
>>
>> Thanks for you help.
>>
>> *-- Russ Abbott*
>> *_____________________________________________*
>> ***  Professor, Computer Science*
>> *  California State University, Los Angeles*
>>
>> *  My paper on how the Fed can fix the economy: ssrn.com/abstract=1977688*
>> *  Google voice: 747-*999-5105
>>   Google+: plus.google.com/114865618166480775623/
>> *  vita:  *sites.google.com/site/russabbott/
>>   CS Wiki <http://cs.calstatela.edu/wiki/> and the courses I teach
>> *_____________________________________________*