Checkout the Showcase of Select Finals (made public with student permission).
- May 1, 2020: the submission site doesn't support multiple partners, but you're allowed to have them, so just click "Ignore Errors" when submitting. We'll make sure everybody gets their grade based on the proposals you submitted.
- May 2, 2020: try to keep your .zip under 1 MB if you can, but we've bumped the limit to 10 MB, so don't worry if you're a bit over
The final "exam" is really an open-ended analysis project. Hopefully it's fun to be free of the usual constraints! You'll be building a static website (no flask) to share your analysis on a topic of your choosing. You'll write Python code (in .ipynb or .py files -- your choice) to generate plots as image files to use in the site.
Being a good data scientist requires you to be both a detective and a teacher. You need to discover truths about your topic, then effectively share them with your audience.
Detective: You will only include 10 visualizations on your site, but your project won't be very good if you only create 10 plots in the process. You should generate dozens of plots for yourself (a loop might generate many plots with the same code snippet!) and invest serious time reflecting on them so that you can learn something about your chosen topic.
Teacher: After you can articulate your most important insights, carefully select 10 visualizations that best communicate the lessons you learned. It's OK (and good) if most of your visualizations get pruned out of the final product. The 10 visualizations in their final form should be highly refined, hopefully not like anything you created when doing your detective work.
The writing that accompanies your plots will be of utmost importance. Read this essay by venture capitalist Paul Graham on useful writing before you get started: http://www.paulgraham.com/useful.html.
Other links:
- Apr 20 [Mon] - proposals due
- Apr 24 [Fri] - send one plot (and corresponding text) to peers for feedback
- Apr 27 [Mon] - peer feedback due back
- May 3 [Sun] - final.zip is due
If you miss the proposal deadline, you will be required to work individually (instead of with a team) for the final.
final.zip will not be accepted late. We'll already be in a time crunch to get final grades out since we can't automatically grade this one. So please submit whatever you have on May 3, even if incomplete.
We recommend forming teams of 2 for the final project. Teams with as many as 4 students total (including yourself) are allowed. Though it's discouraged, you may also choose to work alone.
Consider posting to piazza under the "final-discussion" folder about possible topics to find potential collaborators who share your interests.
You may receive significant help from students on other teams (even sharing code!), as long as (a) they did not choose a very similar topic for their project, and (b) you cite the help you received.
After you submit your proposal, we'll pair you with another team for peer feedback on some initial results you'll share with them (one visualization accompanied by some text). Getting started on a project is the hardest part, so the goal here is to force every team to make that first plot early.
For other CS 320 projects, you are never allowed to post your work publicly online, even after the semester ends. For this final "exam", you are allowed to publish your work after the semester ends, but please don't do so before final grades are released. We don't want teams working on similar projects to copy your code.
Submit it here by Monday, April 20th: https://forms.gle/3DfVmQeHGdPH4yfUA
This lets us know the topic and who will be collaborating on it. Only one team member should submit.
Provides URLs to the data you'll use. You should rely on at least two different data files. There's no upper limit on how many data files you may use! Kaggle is not allowed as a source (others have usually done the hard preprocessing work for those datasets).
Describe the audience for whom you're doing the analysis. You should assume your audience is technical (for example, they've taken CS 320), so this is more about what other background they have and their role in society.
List some of the questions you want to answer and potential actions/policies that will be evaluated. Be sure to tie this to the audience. For example, if the audience is Madison's city council, suggestions related to US federal spending are not relevant.
You site should have four pages: index.html, data.html, acks.html, and results.html (most important).
You may spend time to make the pages aesthetically pleasing if you like, but we only care about the content for grading purposes.
The requirements for each page are given here.
Hand in a final.zip file as P7 at the same place you submit regular
projects:
https://tyler.caraza-harter.com/cs320/s20/submission.html. The zip
should contain your code, pages, and graphics. If the data is small,
include that too (the whole zip must be <1 MB).
Only one team member should submit. The submission site may complain about missing project, submitter, and partner info. The site doesn't support multiple partners (as allowed for the final), so please just check the "Ignore Errors" box this time.
We should be able to perform the following steps to evaluate your work (test it for yourself before handing in!):
unzip final.zipcd finalpython3 -m http.server --bind 127.0.0.1 8000- open
http://127.0.0.1:8000/index.htmlin a web browser
Note: The above is for testing out your submission locally. To do this on your VM, replace 127.0.0.1 by 0.0.0.0 in step 3 and in step 4, replace 127.0.0.1 with your VM's IP address.
Please read the grading standards carefully.
Overall, have fun with this one! Although there are lots of details, keep in mind that your big goal is to present your analysis in a way so that whoever is grading you (or reading your site) learns something interesting in the process.
If you accomplish that big goal, we'll avoid being too nitpicky about the deductions. Learning entails remembering -- what do you really hope somebody who reads your site will remember a month later and will want to tell their friends? Make sure you know what those lessons are, then leverage all your text and visuals to make them stand out!