Applied Data Science (ADS, INFO_I-590, "Topics in Informatics") is an online course first developed by Prof Joanne Luciano and Associate Instructor JT Wolohan and offered in the spring of 2017. ADS was organized around the idealized "classic data science workflow" with emphasis on the real world variability of data science applications, and how appropriate methods and tools should be applied for a given scenario. Tutorials on skills and tools are included (R and Python) but the emphasis is on the application of skills and tools to realistic problems. For the spring 2018 edition of the course, Jeremy Yang joined Kaicheng Yang as AIs for the course. Jeremy was tasked with reviewing the course and recommending and implementing revisions to improve coherence and consistency. Much excellent content was not contextualized sufficiently for some students, who were unclear as to the difference between computational virtuosity and the broader needs of data science. In addition, some updating of the computational and instructional methodology was justified, for example replacing command-line Python with Jupyter notebooks, and replacing R-Commander with R-Studio. Informed by discussions with Joanne and Kaicheng, Jeremy revised the third unit on machine learning algorithms, slightly reducing the number of algorithms covered, but enhancing the contextualization, the understanding of how to select appropriate algorithm classes for given scenarios. Additionally, the "Final Exercise" was revised, to three options: (1) the Kaggle challenge previously required, for advanced students, (2) a Twitter API sentiment analysis problem formerly required in module 9, and (3) a option for an original student proposal. These options enable educational advantages for a range of student profiles.
- Tableau and Cricket Intro
- A8 Tutorial RStudio part 1
- A8 Tutorial RStudio part 2
- Tutorial: Python, Scikit-learn; algorithms Naïve Bayes and Neural Networks
- Twitter API Tutorial part 1
- Twitter API Tutorial part 2
- Why Data Scientists Like Models
ADS Overview Diagram:
Jupyter notebook developed for a video tutorial on the Twitter API final exercise option:
Another addition was to recommend as an accompanying textbook "Doing Data Science" by Cathy O'Neill and Rachel Schutt, which aligns well with the theme and content of ADS. Another addition was a short enriching video lecture "Why Data Scientists Like Models" describing the various important meanings of models and modeling.
Some diagrams and resources related to the workflows of data science:
 |
 |
|---|---|
| Data science teams and roles | Data science flowchart |
 |
 |
| Art of Data Science book | Roger Peng's Data Science Epicycles diagram |