About
This was created by Kelly Ervin Mills as a final project for the MICA MPS Data Analytics and Visualization Program's Design Lab: The Industry Challenge course during the Summer 2023 semester.
Inspiration & Goal
I live in an extremely under-resourced school district in rural
Appalachia. In college, I volunteered at the local Head Start center
and was struck by how much of a difference early educational
intervention can make for children with varying disadvantages who are
struggling to keep up with their peers. I wanted to expand that line
of thinking into K-12 schools and create a piece that can draw
attention to those who need it most.
In an interactive “scrollytelling”/visual essay format, I wanted
locate and draw attention to school districts that likely have the
most students who have fallen behind and are struggling educationally.
Audience
My primary audience is constituents in school districts with the most at-risk students. I believe this is useful for anyone interested in K-12 education in America. But, I am specifically targeting people living in the most affected school districts. I want the stark contrast between the highlighted districts and the rest of the country to hit home for those who call those places home. I hope that the realization that their communities are in need will compel members of my target audience to answer the call to action by contacting their representatives to advocate for funding for their local students.
Data
The data fueling this project is The Urban Institute's " Household Conditions by Geographic School District " dataset. From The Urban Institute: "This dataset describes the share of households within a geographic school district with conditions that may affect remote K–12 learning environments, such as linguistic isolation, crowded conditions, and access to a computer and broadband internet."
Process
I started by loading the Urban Institute dataset into a
Jupyter Notebook
and used the Pandas library to clean and pare down the data. Then I
imported a supplemental dataset to add County FIPS identifiers for the
Urban Institute data. I removed duplicate counties, which may have
led to inaccurate conclusions about counties containing multiple
school districts. A Pandas GroupBy operation and a proportional
average of the school districts making up each county would remedy
this. Time constraints for this project did not allow for this
correction, as GroupBy can be finicky.
At this point, the data required to fill in choropleth graphs was
exported to a CSV. That CSV went into an
Observable Notebook
where I used Observable Plot to create a county choropleth of each a
ttribute from the Urban Institute data.
I exported the choropleths into Canva and manually added callouts to
highlight counties of interest and ensure the maps all lined up
correctly. All these Canva assets went into my site, which uses
Scrollama.js for the sticky scrolling functionality. The rest of the
work was basic web development.
Evolution
This project has evolved a great deal from its inception. After
submitting my concept pitch for review, the feedback that stood out
most was that the project was "ambitious." I dreamed of creating a
large interactive scrolly-telling exploratory piece with many moving
parts. That would have involved reusing some old code but also a lot
of piecing things together and wrestling with web development. After
that feedback, I decided I needed to work hard to not get in my way.
I wanted to focus less on coding a bunch of stuff and more on refining
what I did have. I cut out the time-based component of my visualization
and removed the interactivity that would've been a reuse of a previous
project. The narrower scope, I feel, has allowed me not to bombard the
user with information and exhaust them - something my original concept
definitely would've done. This final concept keeps the story digestible
and the user fresh until the call to action at the end. I believe not
forcing the user to get lost in the weeds serves the goal of having
users take action to contact their representatives.
My audience has also evolved a lot since starting this project. I
originally wanted to target "the American public and especially
policymakers who have the power to direct resources and funding to
school districts in need." This audience was too poorly defined, and I
felt directionless in my design. I pared it down to target residents in
counties in need. Narrowing the audience and creating a specific call
to action instead of just asking people to care about and explore the
issue gave the visualization much more purpose and direction.
This isn't the most technically ambitious project I have completed.
However, I have grown a lot through this process. I am a software
engineer, and coding is just so much fun for me that I get carried away
seeing what I can create until I have a big mess that is exhausting to
wade through. Learning to exercise restraint in my design is my most
significant accomplishment of this project, and I feel like it is the
cleanest story I've told.
Accessibility
While this project isn't 100% section 508 compliant, some design
decisions were made with visual accessibility in mind. This project
uses the
Open Dyslexic
font throughout to increase readability. All charts use single-color
scales that are equally visibly accessible to people with all types of
color-blindness. The single-color scale is also less stimulating for
some neurodivergent folks. Font and background color combinations all
meet to WCAG AA contrast requirements.
Additional changes can, and should, be made to make this project
accessible for a wider audience. Due to time limitations, all charts
are static assets, including the callouts. However, all the callout
text is duplicated in the scrolling text. The charts require detailed
alt-text descriptions to make sense for screen readers. All efforts
should be made to adjust the font/background color combinations to be
WCAG AAA compliant instead of just AA.