CS 4984 - Social Computing Capstone
Fall 2017

Practicum spotlight (10%) - Group

We are going to engage in a series of in-class student driven practicum presentations called spotlights. A spotlight is an opportunity to share a compelling aspect of the social computing capstone project that you are currently working on.

Your practicum must include a technical component which can provide “key insights” into either a data or design aspect of social computing. For example, it can be a feature or library that you are currently using for your project or a brief exploration of a dataset that you found or walkthrough of how you fetched data from a social site. It could also be a design component that you have just built. For example, a browser extension that integrates with Twitter to highlight posts coming from news websites.

During class you will have 10 to 15 minutes to introduce us to your practicum topic. Your goal is that by the end of class, any of us should be able to take your supporting materials and get up to speed rapidly. The overall aim of the practicum spotlight is to allow rapid knowledge and material sharing among students, which can help toward building key components of the final group projects.

Topic sign-up & sign-off

We will hold our sign-up sessions in the first couple of weeks of class, after which you will have an assigned spotlight date. You should email me your spotlight topic by Tuesday (10/3) before class.

Presentation materials (due by noon on the day of your spotlight)

You should send me your materials by email by noon on the day of your spotlight. Also upload on Canvas. For data analysis type presentations, this could be Jupyter notebook, R scripts, the PDF of your slides, etc. For design type projects, this could be a demo, javascript, greaskemonkey scripts, browser extension plugin, etc. Additionally, send the slides of your spotlight. Subject of the email should be - CS4984 Spotlight-[your Team Number]-[your Team Name]. All materials should be sent as a single zipped file. Please name the file as CS4984 Spotlight-[your Team Number]-[your Team Name]. The topic name corresponds to a short identifiable name of what you are going to present in class. I will post the materials in Canvas under the Spotlights folder in the Files tab section. Additionally, I will create a discussion tab in Piazza for your corresponding materials. Your classmates can direct questions to you through Piazza.

In-class presentation

During class, you will have 10 to 15 minutes to present your spotlight topic and quickly walk us through your code-base. I highly recommend that you provide us with a preview of your talk at the very beginning, rather than directly jumping into the implementation details. Here is a good structure for preparing your talk:

  • Introduction / Talk outline
  • What you are trying to do?
  • Why?
  • What will you show us?
  • Then dive deep into the topic, implementation details, how it can be used by others, etc..

Potential spotlight topics that you can choose from.

Feel free to suggest anything else that you think will be useful components for your project. Discuss with me so that I can sign-off on your chosen topic.

  • Different types of methods from scikit-learn. Examples:
  • Data preprocessing using scikit-learn or nltk, pandas.
  • Implementation using NumPy and/or SciPy
  • Web mining using python Pattern
  • Visualizations in Seaborn or any other cool tool.
  • Visualize this by Nathan Yau.
  • Plotting using python Plotly, R ggplot
  • Fetching data through APIs (Twitter - python tweepy, Facebook, Reddit, etc.)
  • Web crawling and html parsing using Beautiful Soup, Scrapy, etc.
  • Text analytics with nltk or with spaCy or with other off-the-shelf tools, (e.g. showing a script or tool to do parts-of-speech, named entity recognition etc.)
  • Browser extensions (Chrome, Firefox) to modify an existing site (e.g. change affordances for Facebook comments, Twitter replies, Reddit responses etc.). Greasemonkey and TamperMonkey are useful user script manager to dig into. Here is a very WELL DONE google chrome extension - PolitEcho that finds out how polarizing the content on your Facebook news feed is when compared to your friends as a whole. Check it out!

Grading of spotlights

For each spotlights, you grade will be assessed as follows:

(10%) Topic sign-off

  • Did you submit your topic to me on-time? Late submissions will receive a 0 on this component.
  • Did you put care into your topic choice? Did you make a convincing argument as to why this is a good spotlight topic?

(50%) Presentation materials

  • Did you submit your materials on-time? That is, by noon on the day of your spotlight? Late submissions will receive a 0 on this component.
  • Are your materials well-written?
  • Is your code clear and well-documented?
  • Did you provide clear examples?
  • Did you leave out important issues?

(40%) In-class presentation

  • Preparation and Poise: Were you well prepared to present to the audience? Did you engage with them?
  • Use of visual aids: Did you use pictures/diagrams to explain your ideas? Did the slides contain short, clear bullets rather than long sentences and/or cryptic equations?
  • Answering questions: Did you address technical questions and comments well, including questions on Piazza?