35 Years of American Deaths

By Jason Guo

Link to visualization website: https://projects.fivethirtyeight.com/mortality-rates-united-states/#2014

How was the data collected?

This data was gathered from the Institute for Health Metrics Evaluation and they designed a statistical model that uses demographical and epidemiological data to assign specific causes to records containing codes in the National Vital Statistics System, which gathers death records from state and local jurisdictions into a national database. This data was then all brought together and organized by the author of this data visualization, Ella Koeze.

The selection bar can be changed to see deaths relating to specific causes such as cardiovascular diseases or interpersonal violence. If there are different categories relating to a certain cause of death, then it could be isolated to show just that cause of death specifically.

The author of this visualization also offers some charts pertaining to the counties with the lowest and highest mortality rates in the United States during the year of 2014. This was particularly cool to summarize the rates on the far left and right ends so that the user does not have to spend time comparing each county on the map. Some of the counties have high mortality rates because they are near the Rural Appalachia because of rising cancer rates and deaths due to substance abuse. Other counties such as the ones in North and South Dakota have high mortality rates because of the areas made up of mostly American Indian reservation lands because historically people there suffer from poor health outcomes and challenges in healthcare access. As for the counties with lower mortality rates, many of these are sparsely populated and others are really wealthy such as Fairfax county, Virginia and Los Alamos county, New Mexico.

Pros

I like how the map is color coded in gradient form of same color to emphasize the higher or lower deaths in certain counties, the darker the color, the more deaths there are and the lighter the color, the less there are. This can be beneficial to people with colorblindness because no comparison of 2 different colors are needed, only the shade of each color.

The above 3 pictures show the different color coding for <10 deaths per 100k people, <100 deaths per 100k people and >=100 deaths per 100k people. Not only do the color gradients matter, the different category of color in general shows the user a feeling that the certain types of deaths occur way more than others on the 100k scale. There is also a legend to help differentiate the color shades.

On a the US map only the states are easily defined and counties are very small and some can barely be seen so we can zoom it in to look at the county that is clicked and hover the cursor over the counties to show the average deaths per 100k for each county, state and United States.

Cons

• Since the counties are small and the user wants to view a county more clearly, the user would want to zoom into a county and the surrounding areas, however the user can only click to zoom one time and then clicking again would automatically zoom back out. There is also no functionality to click and drag with a cursor to browse the map in the zoomed state, the user would have to zoom out then click on another area to zoom in again. Both of these are shown above.
• The overview chart that the author gives for most/least deaths for certain counties is not really dynamic meaning that the year could not be adjusted to look at the patterns or changes that occur over the years 1980-2014 and see if Union county always had the highest mortality rate or if it was another state, it was only for 2014, so if more interactivity were to be added to this area, it would be great.
• One problem in particular that I saw is that when the deaths are recorded, it is from where the person is from and not where they died. So if a person is from Idaho, traveled and died in Alaska, the death is recorded in Idaho. This can skew data especially if something from a different state killed the person like murder or a disease.
• The legend and data is consistent and all but the "deaths per 100k people" label is not accurate for comparing big and small states because of the population that resides in these states. Some are small and saturated like counties in New York and some are big and unsaturated like counties in Colorado. Things like putting percentages of mortality rates per population can clear up some confusion in the population gaps because the users are used to looking at things relating to proportions to a population.
• Even though the visualization can show the averages per county while looking at the averages for the state and country, There is not option to compare county by county together. I think this is a huge part missing from the visualization because if I want to look at a county in California and one in New York, I cannot compare them side by side. There is an attempt for me trying to compare information from 2 counties shown in the GIF below.

Who are the users that this visualization was made for?

The target audiences for this visualization are students, anyone interested in the causes and amount of deaths occuring in the United States and the general public.

What questions do people want to ask?

• What state or county had the least/most deaths?
• What are the causes of the least/most deaths?
• What year did the doctors find a cure or treatment for a disease so the deaths declined?
• What year did a disease start and spread across the nation?
• What kind of disease are people mostly or least likely to die by?

How can they find the answers with this tool?

• What state or county had the least/most deaths? Depending on the year the user wants, the user can use the slider to select the year then look at the lightest shade of the color(could be white) on the map and record it as the county/state with the least death and the the darkest shade on the map and record it as the county/state as the most death. If the colors are similar, the user can hover over the counties individually to compare averages rates.

• What are the causes of the least/most deaths? We can find the causes of the least and most deaths by selecting the drop down menu and choosing the causes on the top and the bottom of the list. This is because they are sorted by the most causes of deaths on top by descending order. This is shown above.
• What year did the doctors find a cure or treatment for a disease so the deaths declined? Let's say for example we want to find the year when we want to find the year of which a doctor found a treatment for HIV/AIDS. We would see the decline of the mortality rate of the disease during the years that follow meaning that the US map would get lighter. In this visualization it shows that the treatment was found and used from 1996. This is shown on the bottom right GIF.
• What year did a disease start and spread across the nation? We can talk about HIV/AIDS again, so since it was not a native disease and we know that it was brought into America, we can see this in effect when one point of the US map gets darker. In this visualization we can tell that HIV/AIDS started to spread form Florida. This is shown on the bottom left GIF.