Pet project

What is a silly/quirky/fun/interesting thing in your own life or environment that you could datafy and analyze?

Data mining gives you a powerful toolbox, and I want to give you a chance to use it in something that is of interest to you.

The course includes a personal project in which you should work throughout the trimester. It should be something that has some personal connection to you. This is neither about downloading an existing dataset and working with it, nor about participating in a canned data science competition.

Instead, this is about creating and analyzing a personal dataset about something that is connected to you, as an individual. It can be something silly, quirky, or fun, as long as it is interesting, let me give you some examples:

• Keep track of a health parameter (e.g., mood, exercise, etc.)
• Keep track of your money expenses (all of them, or specific categories such as entertainment or food).
• Keep track of your food intake.
• Keep track of all the ads you see in the street.
• Download data from a connected device that you own (e.g., steps, heartrate, sleep, etc. from a smart watch), or better, from multiple devices.
• Download data from a service you use (e.g., trajectories in a map service, music or video history in a streming service), or better, from multiple services.
• … and so on, these are just examples.

In your analysis, you can use techniques we have seen in class, or other techniques.

Think of a fun “pet project” worthy of a data scientist. Make it personal, interesting, fun, and wholesome.

Pre-delivery: concept note (1 slide)

You should pre-deliver a concept note. One slide, with readable text, containing:

• A descriptive title, e.g.: “Sleep analysis of domestic cat” followed by an optional evocative title, e.g.: “The purr-fect nap”
• (A) How will you capture data, e.g.: “I will stay with my cat during her naps”
• (B) How much data will you capture, e.g.: “I will capture a minimum of 20 naps”
• (C) Which features will you capture, e.g.: “I will capture: timestamp, eye movement, paw movement, whisker movement, and time slept/aware”
• (D) Which question will you answer, e.g.: “I have the suspicion my cat sleeps more than other cats, I will compare against the average of 15-30 minutes. I also want to know if my cat rests, by checking how much she moves during sleep.”

The ideal project answers a specific question, and has an actionable and specific outcome. Do not submit something vague like “understanding” or “exploring”.

Delivery: report (3 pages)

A three-page report using ACM or IEEE proceedings format, including:

• Title: should be descriptive or fun (puns encouraged)
• Goals (1-2 paragraphs): what are your motivations and objectives, and how is this connected to you as an individual.
• Methods (0.5 pages): how you created and processed your dataset.
• Results (1.5 pages): what did you observe, using 2-3 (but not more!) tables or figures as support but not subsitute of your narrative.
• Conclusions (1-2 paragraphs): a closing paragraph with your learnings and thoughts, and what does it mean to you personally.
• Generative AI Statement (1 brief paragraph): a paragraph in which you indicate whether you used Generative AI to do your project, and how, specifically.
• Repository (1 link): a link to a repository containing your code and data
  • It should be professional, understandable, and reproducible.
  • It should have a good README.md file in the root.
  • This repository may be private or public, as you prefer.
  • If you use a private repo on GitHub, share it with GitHub user chatox.
  • If you use a private repo elsewhere, share it with the e-mail of the instructor.

Your project belongs to you. I will not distribute it.

How it will be evaluated

The personal project is worth 30% of your practice grade. It must be:

• interesting or fun,
• non-trivial, and
• perfectly executed.

Please note that the following will lead to a zero grade.

• Your project has nothing to do with you, e.g., you simply used an existing dataset.
• Your report does not look as if it was written by you. This is about you, I want to read your writing.
• Your project captures other people’s personal data without their consent. Nobody else but you should be identifiable by your report or your dataset.