Main Project (Group Submission)
-
Grading is subjective: Insight over charts/tables
-
Rubrics are weighted according to appropriate project context
For example, project where clustering is inappropriate and data fitting is
critical, points move from data fitting to clustering
-
+ indicates: must have (if appropriate, see previous point)
* indicates: extra points (desirable, but do not try to artificially introduce)
-
data:
-
+ characterize: incomplete, uncertain, outliers
-
+ filter/scrape/clean (complexity gets * )
-
* large data sets
-
* streaming data sets
-
+ combine different data sets
-
hypotheses
-
+ clearly stated
-
+ devise numerical experiment(s) to confirm or deny
-
* causality vs correlation
-
fit data
-
+ regression, MLE, learning (increasing points)
-
* prediction
-
other data science techniques
-
+ cluster data
-
...
-
presentation.
-
* engaging visual representation of data
-
+ conclusions
Submit:
-
supporting material that allows us to verify your work: code or worksheet(s), overview, url or other means to access the data you used
-
presentation material: 4 minute video presentation of your hypothesis and findings based on data science techniques.
(hint: see the grading sheet below)
Use zoom or youtube and send the url so I (not you) can start it in class.
-
A
grading sheet
(editable)
-
A writeup of 2 pages or less (verbosity is not a criterion) that allows you to explain or expand on points that
may be unclear when viewing just (b) and (c).
Presentation:
-
4 min video
-
group will come to the black board during the video
-
2 minutes questions