For your final project, you will take a dataset, explore it, tinker
with it, and tell a nuanced story about it using at least three charts.
I want this project to be as useful for you and your future career as
possible - you’ll hopefully want to show off your final project in a
portfolio or during job interviews. Accordingly, you get to choose what
data you will use for your project. Choose whatever one you are most
interested in or will have the most fun with.
Overview
For the final project, you will:
- Write a research question.
- Conduct an exploratory data analysis of one or more data sources to
address your research question.
- Create visualizations to effectively communicate the results of your
analysis using the principles of information visualization covered in
this course.
- Document your entire analysis in a reproducible format, including a
detailed curation of the data used, analysis conducted, and results
produced.
Click here for a
template to use for your analysis.
To make the overall project more manageable, it is broken down into
separate “milestone” deliverables due throughout the semester, starting
with a proposal due just before spring break.
Summary of Deliverables:
| Proposal |
6 % |
March 13 |
| Progress Report |
8 % |
April 22 |
| Peer Review |
5 % |
April 29 |
| Final Report |
16 % |
May 04 |
| Presentation |
8 % |
May 06 |
Purpose: The skills covered in this course are
rooted in design rules and principles rather than formulas and
equations. As such, the application of these principles to a
real data problem is one of the best ways to learn and assess mastery of
these skills. I guarantee you one day you will need to apply these
principles to communicate an idea to colleagues, so let’s make sure you
have at least one chance to practice before the stakes
are higher.
Skills & Knowledge: Your final project should
illustrate your ability to transform raw data into insights by making
the non-visible visible, showing clear trends or patterns, and / or
identifying outliers or missing information. The specific skills
involved in achieving this goal include all of the course learning
objectives listed on the course home page.
Teams: You may work on an individual project or in
teams of 2 to 3 students of your choosing. If you choose to work in a
team, you must finalize your team members at the time of submitting your
proposal.
Proposal
Due: March 13 by 11:00 pm
Weight: This assignment is worth 6% of your final
grade.
Submission Details: You may write your proposal
using any tool you wish (e.g. Microsoft Word, RMarkdown), but you must
submit your proposal as a PDF document on Blackboard by
the due date.
Assessment: We will use this rubric to grade
your proposal.
Tasks:
Write a 1-2 page proposal of an exploratory data analysis that you
plan to conduct for your final project. The instructors will review and
grade your proposal and provide feedback upon returning from spring
break. If your proposal is approved, you’re done and can move on towards
the next project task. Otherwise, the instructors may ask you to submit
a revised proposal, most likely by focusing / adjusting the proposal
scope and / or research question. Below is a list of specific items your
proposal should include (check the rubric to see their relative
weighting).
Follow these formatting guidelines:
- Be between 1 and 2 pages in length.
- 12 point font, 1 inch margins.
- Write your project title at the top of the document
(e.g. Title:
<insert title>).
- List the name(s) of student(s) involved in the project below the
project title (e.g. Authors:
<insert list of names>). You cannot change your
teammates once you’ve submitted your proposal.
State a clear research question. Follow these
guidelines. Your question should be:
- Clear: it provides enough specifics that your
audience can easily understand its purpose without needing additional
explanation.
- Focused: it is narrow enough that it can be
addressed thoroughly with the data available and within the limits of
the final project report.
- Concise: it is expressed in the fewest possible
words.
- Complex: it is not answerable with a simple “yes”
or “no,” but rather requires synthesis and analysis of data.
- Arguable: its potential answers are open to debate
rather than accepted facts.
Discuss the data source(s) you plan to use for your analysis:
- If you have already identified the source(s), describe them and
include urls and / or references to the sources. If you have not
identified a source yet, describe the data you hope to use, and give at
least one plausible source that may have the data (regardless of whether
the source makes it available or not).
- Discuss the validity of the source(s). For each data source, is the
data available the original data, or has it been processed? How
was the original data collected and by whom? If you do not yet have a
source, discuss what concerns you have about a plausible source that
might have the data.
Describe some anticipated results:
- Choose two variables that you expect to find in your data and are
relevant to your research question. Describe how you would expect each
to be distributed (e.g. unimodal, multimodel, tightly-group,
widely-spread, etc.), regardless of whether those variables actually
exist. For example, you might expect the price of gasoline over a
particular period to be rather tightly-grouped around a mean, whereas
the stock price of a particular company might vary much more widely over
the same period.
- Choose two or more variables that you expect to find in your data
and are relevant to your research question. Describe two relationships
you expect to find these variables, regardless of whether actually exist
in your data. For example, you might expect sales of hybrid vehicles to
increase when gas prices increase, or sales of SUVs to increase when gas
prices decrease.
- Describe at least two charts that you expect will help you visualize
the relationships that you expect to find. For example, a scatterplot of
gasoline prices and SUV sales over a particular time period might be
useful for visualizing the level of correlation between these two
variables.
- Discuss how your expectations about the variables you chose will
help inform you about your research question.
Progress Report
Due: April 22 by 11:00 pm
Weight: This assignment is worth 8% of your final
grade.
Submission Details: Use this template for your analysis and
report. Your progress report should be written as a .Rmd file
that compiles to a html webpage. Submit your entire project folder
(including your .Rmd file, data files, .RProj file, etc.) as a single
.zip file on Blackboard by the due date. For students in teams,
only one person from your team should submit the
report.
Assessment: We will use this rubric to
grade your progress report.
Tasks:
Write a report summarizing progress you have made towards your
project thus far, including summary statistics of your data and
preliminary charts. You should have already identified data source(s)
for your project and begun exploring the data. Students working
in teams must submit individual short reviews of their teammates’
contributions to help ensure that the workload and grading for
team members are equitable (see task 6 below). Below is a list of
specific items your progress report should include (check the rubric to
see their relative weighting).
Follow these formatting rules:
- As the report will compile to a html webpage, there is no length
requirement; your report should be sufficiently detailed to address the
requirements listed below and sufficiently concise such that it is
expressed in the fewest necessary words.
- In your markdown YAML header, include the project title and the
name(s) of student(s) involved in the project such that they appear at
the top of the rendered html page.
- Your report should be written in a narrative format (i.e. using
coherent pargraphs rather than a series of bullet points). You may use
headings where appropriate to break up your report into sections.
State a clear research question. This can (and almost certainly
should) be a revised question from your proposal. Again, follow these
guidelines.
Discuss the data source(s) you are using for your analysis:
- Describe them and include urls or references to the sources.
- Discuss the validity of the source(s) - for each data source, is the
data available the original data, or has it been processed? If
not, what is the original source? How was the data collected?
- For each dataset, provide a data dictionary in an appendix at the
end of your report that contains a table of each the name, type, and
description of each variable. The appendix does not count towards your
page limit.
Evaluate your proposal expectations:
- Do you find support for your original proposal expectation about how
one variable might be distributed? Show it by looking at summaries of
the variable. If that variable doesn’t exist in your data, discuss
another key variable instead and how it is distributed.
- Do you find support for your original proposal expectation about how
the variables you chose are related? Show it. If those variables don’t
exist in your data, discuss one key relationship that you have
identified between two or more variables.
Include at least two preliminary charts:
- Your charts should either support or oppose your research question,
or they should illustrate what else you might need to address your
research question.
- You may choose whatever chart types you wish, but your choices
should highlight the point you want to make or should clearly show the
relationship you want to emphasize. Consider these resources when choosing your
charts.
- Your charts should follow the design principles we have covered in
class. They do not have to be fully polished yet, but at a minimum they
should be accurate (i.e. not misleading) and they should not include
distracting non-data ink.
For students working in teams: On Blackboard
under the assignment titled
"Team Member Review [Progress Report]", submit a short
description of the specific contributions of each team member in your
team. For example, “Student A identified the data source and wrote the
documentation for it. Student B led the data cleaning process and did
much of the initial data exploration. Student C helped write code for
the main visualizations.” These reviews will be kept confidential and
compared to assess that the workload and grading for team members are
equitable. Team members who do not make meaningful contributions to
their projects will receive a lower grade than that of their team mates.
If you are having any disputes amongst team members, please
contact Professor Helveston so we can find a
resolution.
Peer Review
Due: April 29 by 11:00 pm
Weight: This assignment is worth 5% of your final
grade.
Submission Details: Submit your peer reviews as a
PDF document on Blackboard by the due date.
Assessment: Your peer review will be graded for
completion.
Tasks:
Each student will be randomly assigned one progress report of their
classmate’s projects to review. This is an individual
assignment, so students in teams must individually write their
own reviews. Students in teams will each be assigned different projects
to review. The reviews will be returned to the project teams as further
feedback to consider as they revise their analyses and prepare for their
final report and presentation. Below is a list of specific items your
peer review should include (check the rubric to see their relative
weighting).
Follow these formatting rules:
- Be between 1 and 3 pages in length, including all images.
- 12 point font, 1 inch margins.
- Write the title of the project being reviewed from the progress
report (e.g. Review of:
<insert project title>)
- Do not put your name on your review - these will be
anonymous reviews returned to each team.
Evaluate their research question according to these
qualities:
- Clear: it provides enough specifics that your
audience can easily understand its purpose without needing additional
explanation.
- Focused: it is narrow enough that it can be
addressed thoroughly with the data available and within the limits of
the final project report.
- Concise: it is expressed in the fewest possible
words.
- Complex: it is not answerable with a simple “yes”
or “no,” but rather requires synthesis and analysis of data.
- Arguable: its potential answers are open to debate
rather than accepted facts.
Evaluate the description of their data source(s):
- Are the data appropriate for the research question?
- Are the data sources well-described? Do you know the source and the
original source?
- Can you follow the information in the report to obtain the data
yourself?
- Is there a data dictionary available in an appendix for each data
set?
- Provide one suggestion to help further clarify any questions you
have about the data source(s).
Evaluate their preliminary results:
- Are there at least two charts provided in the analysis?
- Are the charts relevant to the research question posed?
- Do the charts support or oppose the research question or illustrate
what else might be needed to address the research question?
- Do the charts follow the design principles (to a reasonable degree)
that we have covered in class (i.e. they are not misleading, and they do
not include distracting non-data ink)?
- Provide at least two suggestions to improve the charts.
Final Report
Due: May 04 by 11:00 pm
Weight: This assignment is worth 16% of your final
grade.
Submission Details: Use this template for your final analysis
and report. Your progress report should be written as a .Rmd
file that compiles to a html webpage. Publish your compiled page online
(e.g. via RPubs, Github, etc.), then submit your entire project folder
(including your .Rmd file, data files, .RProj file, etc.) as a single
.zip file on Blackboard by the due date. Also include a link to the
published report page in your Blackboard submission. For students in
teams, only one person from your team should submit the
report.
Assessment: We will use this rubric to
grade your report.
Tasks:
Your final report should be a fully reproducible product and
available online as a html webpage. It should include text, data, code,
and plots. Students working in teams must submit a short review
of their teammates’ contributions to help ensure that the
workload and grading for team members are equitable (see task 6 below).
Below is a list of specific items your report should include (check the
rubric to see their relative weighting).
Follow these formatting rules:
- As the report will compile to a html webpage, there is no length
requirement; your report should be sufficiently detailed to address the
requirements listed below and sufficiently concise such that it is
expressed in the fewest necessary words.
- In your markdown YAML header, include the project title and the
name(s) of student(s) involved in the project such that they appear at
the top of the rendered html page.
- Your report should be fully reproducible - all data
formatting and charts should be written in code chunks and rendered when
you compile your .Rmd file to a html webpage.
- Your report should be written in a narrative format (i.e. using
coherent pargraphs rather than a series of bullet points). You may use
headings where appropriate to break up your report into sections.
- Proofread your html webpage before you submit - double check for
spelling and formatting errors, especially rendered charts and
tables!
State your research question and motivate why it is important /
why the reader should care.
Describe your data source(s) you used for your analysis:
- Describe them and include urls or references to the sources.
- Discuss the validity of the source(s) - for each data source, is the
data available the original data, or has it been processed? If
not, what is the original source? How was the data collected?
- Provide a data dictionary in an appendix at the end of your report
that contains a table of each the name, type, and description of each
variable in your data. Include a dictionary for each dataset used.
Describe your data:
- Articulate the main variables of interest in your project, and
justify your choice of variables.
- Provide descriptive statistics for your relevant variables. These
can be a mix of graphs and summary tables.
- You don’t have to summarize everything in every data set - just the
variables that are relevant to your analysis.
Describe your results:
- Display charts that either support or oppose your research question
or illustrate what else you might need to address your research
question.
- Write narrative text around your charts to explain what the charts
show and their significance towards addressing your research question.
This should read as a continuous story rather than as a reply to each of
the requirements described here.
- Your plot type choices should highlight the main point(s) you want
to make or clearly show the relationship you want to emphasize.
Basically, they should answer the question “what do the data say about
my research question?”
- Your charts should be polished, following the design principles we
have covered in class.
- You must include at least three polished charts / animations;
alternatively, you can build an interactive shiny app with at least two
dynamic charts. If you build a shiny app, include a link to your app and
screenshots of your app in your final report.
For students working in teams: On Blackboard
under the assignment titled
"Team Member Review [Final Report]", submit a short
description of the specific contributions of each team member in your
team. For example, “Student A identified the data source and wrote the
documentation for it. Student B led the data cleaning process and did
much of the initial data exploration. Student C helped write code for
the main visualizations.” These reviews will be kept confidential and
compared to assess that the workload and grading for team members are
equitable. Team members who do not make meaningful contributions to
their projects will receive a lower grade than that of their team
mates.
Presentation
Due: Slides due by May 05 by 11:00 pm; presentations
will be in class on May 06.
Weight: This assignment is worth 8% of your final
grade: 4% for your slides & slide design, and 4% for effectively
presenting them in class.
Submission Details: Submit your slides as a PDF or
.pptx file (whichever you feel most comfortable presenting with) on
Blackboard by the due date.
Assessment: We will use this rubric to
grade your presentation.
Tasks:
You will present your analysis with your teammates in class on May
06. Your slides do not need to include the code used to conduct your
analysis (that should be able to be obtained from your report). Instead,
your presentation should be a short explanation of 1) what you studied
and why it matters, 2) what data you used, and 3) what you found. Below
is a list of specific items your presentation should include (check the
rubric to see their relative weighting).
Follow these formatting rules:
- Your presentation should be no longer than 10 minutes in length
(practice! time yourself!)
- Each team member must present at least one slide.
- You should have between 6 and 8 slides total, including your title
slide.
- Your title slide should include the project title, team member
names, and the presentation date (May 06, 2020).
- All slides should be numbered in the bottom-left or bottom-right
corners.
State your research question and explain why we should care (you
don’t necessarily need a slide with the question written out on
it).
Discuss the data source(s) you used for your analysis:
- Describe them and display the names of the sources on your slide.
You do not need the full urls to the raw data files, just the
organization name
(e.g.
"**Data source**: U.S. Centers for Disease Control and Prevention").
- Discuss the validity of your source(s) - for each data source, was
the data available the original data, or has it been processed?
If not, what is the original source? How was the data collected?
Describe your data:
- Articulate the main variables of interest in your project, and
justify your choice of variables.
- Provide descriptive statistics for all your relevant variables.
These can be a mix of graphs and summary tables.
Describe your results:
- Display charts that either support or oppose your research question,
or illustrate what else you might need to address your research
question.
- Your plot type choices should highlight the point you want to make
or clearly show the relationship you want to emphasize - basically, what
do the data say about your research question?
- Your charts should be polished, following the design principles we
have covered in class.
Examples
Here are some examples of exploratory data analyses:
- Is
Usain Bolt really the fastest man on earth?, by Anindya Mozumdar. This analysis
uses some statistical modeling (which won’t cover), but the use of
visualizations alone are effective for addressing the research questions
posed.
- F1
racing data analysis, by Jonathan Bouchet. This is an extensive
exploratory data analysis of a variety of data on races, drivers, and
constructors across the history of Formula 1 racing. This would
probably be overkill for a course project, but sections of it
(e.g. section 3 on the races, or sections 4 and 5 together on the
drivers and constructors) would make a nice project. It involves
mutliple data sets, some not-so-simple data cleaning, and a wide variety
of nice visualizations.
- Increasing
Intensity of Strong Atlantic Hurricanes, by James B. Elsner. This is
a short analysis with the goal of validating the predictions of this study
published in Nature in 2008, which predicted that Atlantic tropical
cyclones would continue to get stronger on average over time. While this
analysis is too short to be considered adequate for a final project, it
is still a good example of good data source documentation and one
effective visualization. An expanded inquiry into the questions
addressed here would make a good course project.
- Death:
Reality vs Reported, by Owen Shen: This is an interesting original
analysis in which the author collected the data himself. The analysis
compares whether the proportion of actual causes of death is consistent
with the proportion of death causes that the media reports. This
analysis also uses some interactive graphics so that the reader can view
different versions of the same plot, such as scrolling through the plot
over time.
- Are
first babies more likely to be late?, by Allen Downey. This is
another relatively short analysis, but the author does a good job of
documenting his data sources, stating the assumptions he makes, and also
describes what data were dropped from the analysis (i.e. babies born via
C-section).
Page sources:
Some of the content on this page is inspired by and / or modified
from other sources, including: