The goal of our project is to provide a visualization of femicide data. The data we visualize has been collected in a political context and can be labelled as “counter data” meaning that it answers a need to reveal a phenomenon usually hidden or marginalized.
We think about three types of personas for our website. First, the people concerned by femicide because one of their relatives died from it. For this user, our goal is to provide a virtual “memorial” and show that their loved ones are not forgotten. The portrait view design follows this goal focusing on the women data and providing an emotional design. The link to other views allow to contextualize these brutal deaths in a larger issue. We would like to add a resource page to provide some option for care in this difficult context.
The second and third personas, are respectively a militant in a search for data to support their cause and a curious mind looking to know more after discovering the topic. For both these personas, the ensemble view as well as the statistical view let them explore the dataset to get an better idea of who the victims are and what we know about them. Also, we would like to add a page with links to the groups that organize the work on the topic of femicide.
Our minimal viable product includes 3 different visualizations linked with each other: portrait view, ensemble view and statistical view (with an exploratory and a aggregation mode).
The sketches for the portrait view of our visualization are presented below. The idea is to create a "memorial" of each victim. The schematic representation can be seen in Portrait View 1: name of the victim is in the center of the circle, with a picture of the victim if it is present in the dataset. The selected year is shown as a large circle (for now we have only data for 2018, if we get more data, a feature to select year will be added to the page). The date of death is placed on this circle and the position is based on the date. A symbol next to the death date denotes the cause of death. Smaller circles show the age of the woman: number of circles corresponds to the number of years. In the right upper corner name of the state is written inside the contour that has a shape of this state on the map. In the left upper corner the information about victim's race/ethnicity is shown (each circle is a race/ethnicity). Portrait View 2 shows another possible way of presenting this information (the race is in a "constellation" design). Portrait View 3 and Portrait View 4 demonstrate other design ideas for the age and year visualization (age is shown as ticks or rays and year is divided into months).
| Portrait View 1 | Portrait View 2 |
|---|---|
 |
 |
| Portrait View 3 | Portrait View 4 |
 |
 |
Here are the sketches for the ensemble view. The starting view (Ensemble View 1) is a plot with as many circles as death cases in the dataset for this year. Below there is a filter by attributes, once a filter is selected the circles become colored based on this filter (Ensemble View 2). If a user clicks directly on a circle he goes to the portrait view of the associated woman. If a particular attribute is selected the graph is colored based on this attribute. The second column next to the filter shows different categories with the distribution (Ensemble View 3). If a user clicks back on race/ethnicity the initial random view will appear. Clicking on the "Total" or on the "Back to overview" leads back to overview. Note that each circle is linked to one case, and that clicking directly on it will lead to the portrait view.
| Ensemble View 1 | Ensemble View 2 |
|---|---|
 |
 |
| Ensemble View 3 | |
 |
| Exploratory Statistical View 1 | Exploratory Statistical View 2 |
|---|---|
 |
 |
| Exploratory Statistical View 3 | |
 |
The goal of the exploratory statistical view is to visualize both the data related to the victims as well as the killers at the same time. The exploratory mode shows all the data, i.e. one line per case and highlights the relationship and the murder details. The same data is also presented in an aggregated way to show the trends in both the victim and the killer data.
| Aggregated Statistical View 1 | Aggregated Statistical View 2 |
|---|---|
 |
 |
The aggregated statistical view shows bar chart for data on women and men sides. For categorical data present for both sets (e.g. women's races and men's races) the graph will be asymmetrical. For categorical data that pertain to both genders together (e.g. relationship) the graph will be symmetrical. For the ordinal data, we will define some categories (e.g. age) for the aggreagation.
In general: We need the first seven lectures to understand the basics of creating interactive visualizations in the browser. Here is a short overview of lectures and tools that we will need.
"Web Development", "Javascript part 1", "More Javascript", "D3.js" - learning about HTML for structuring the webpage, CSS for the styles of elements, SVG for describing graphics, JS and d3.js for creating complex data visualizations.
"Data" - working with the data: exploration, pre-processing, cleaning, analysis, using for data visualization.
"Interactions", "More interactive d3.js" - learning about adding interactivity to our data visualization, very important for all 3 components of our core visualization
"Mark, channel" - creating different elements of our visualizations to represent items and links.
"Perception colors", "Designing viz", "Do and dont in viz" - studying the basics of design process and designing our visualization.
"Storytelling" - one of the most important letures in the context of our project because our main goal is to honor the memory of women killed by men, as well as to highlight the problem of femicide nowadays.
"Tabular data" - for working with a tabular data, creating visualizations for the statistical view.
For getting additional knowledge necessary for completing our project we are going to rely on the following sources: "Interactive Data Visualization for the Web" by Scott Murray and "Visualization Analysis & Design" by Tamara Munzner.
To enhance the visualization in our project, we can implement 2 additional views for our website: the timeline view and the map view (see sketches).
The idea of the timeline view is to present femicide cases in the dimension of time as a story with victims' photographs and names surrounded by women killed around the chosen date. The timeline view is linked with the portrait view by the date of murder element in the portrait view. The timeline can be zoomed in and out in the 2018 year.
Another idea is to create a map view with a map of crimes committed in different states of the USA in 2018. This visualization would help to show some additional insights into the data (states with the largest/smallest number of femicide cases, the distribution in a space domain). To implement this idea, we would also need to use the material from lectures "Maps" and "Practical maps".
Folder with a project prototype
This page is a first glance on the skeleton of the "Ensemble View".
It tries to have an SPA approach for now, but it might later be linked to other pages with a <nav> section if needed.
See there a preview of the skeleton itself.