Change Log for 91-DIVOC #01

This change log lists the major changes made to "An interactive visualization of the exponential spread of COVID-19".
An interactive visualization of the exponential spread of COVID-19 >>

Dec. 31 – Vaccine Data Available for Countries and Detailed Synchronization Logs

Two major additions: more vaccine data (country-level data, instead of just states) and a detailed log of the synchronization status of 91-DIVOC.

Dec. 24 – Vaccine Data Available for US States

As the vaccination against COVID-19 begins, many US states are beginning to report total numbers of vaccinations. This data is now provided on the US state-level graphs, including the normalized views. This data is sourced/compiled by the Centers for Civic Impact at Johns Hopkins University, as JHU continues to lead the way in providing fantastic, open-source COVID-19 data.

The four new 91-DIVOC visualizations include:

Happy Holidays! :)

Dec. 7 – Expanding Graphs on Large Desktop Monitors + 21-Day CFR Experimental Graph

On large/wide monitors, all of the content is centered and there may be a large amount of whitespace on either side of the graphs. Today’s update adds a new option, which you’ll only see on a large monitor, named “Expand Graph”. You’ll find it to the right of “+Add Additional Highlight” on all four of the graphs – click it and see the graph fill up your screen! :)

Additionally, a new experimental graph (21-day lagged CFR) was added at the bottom of the “Data’ selection options.

Dec. 2 – Added Canadian Provinces/Territories

As our friends to the north see their own second wave, several have requested the Canadian provinces/territories added to the 91-DIVOC graphs. To keep the graphs largely the same, I’ve added the Canadian Territories as an “off-by-default” addition:

Nov. 2 – Improving the Y-Axis on Log Scaling

With the introduction of derivative charts, the new code for log scaling on some graphs (particularly those with large values, like a graph of cumulative deaths) left a lot of whitespace at the bottom of the graph due to x-intercept of the y-axis was “fixed” at starting at 0.001.

Log scales are tricky as they are undefined at 0, so you are unable to start the y-axis labeling as 0 and visualizations are forced to choose a non-zero value. (This differs from a linear scale, where best practices dictate that – unless it’s both obvious to the reader and for a clear reason – an honest visualization will always start with the y-axis at zero.) With today’s update, I’ve implemented new code to calculate the starting value (x-intercept) for the log scales.

In general, the scale for all log graphs will start at the 10x power such that 10x is lower than the lowest positive data point, with two exceptions:

Oct. 30 - More Derivative Charts

I received a number of questions (and a lot of interest) on the derivative charts that were recently added – the charts helped show the overall trends, but the day-of-week swings in the underlying data made the chart less useful beyond that. To help provide more context around the derivative charts, I’ve added derivative charts of the “One-Week Rolling Averages” to answer the question: What is the daily change in the 1-week rolling average? Additionally, this new derivative can also be viewed, itself, as a one-week average.

Oct. 28 - Derivative and Experimental Charts

Added several new data selections:

Oct. 18 - New Guide and Normalized Graph Improvements

I’ve added a small guide to the visualization, providing an overview of the key aspects of the visualization including the data sources, regions, normalized data, and more. Additionally:

Oct. 15 - Big Ten School Visualization Colors

Starting a bit over a month ago, Johns Hopkins University has stopped reporting any state-level hospitalization data (you can see the empty column in their raw data). Fortunately, the COVID Tracking Project by The Atlantic continues to report this data and, when choosing state-level hospitalization data graphs, you are prompted to select the COVID Tracking Project data source to view the data. (It appears they’re not tracking every state, but they’re getting data from the majority of the states.)

Additionally, the mobile layout labels were getting extremely cluttered – the latest update cleans up the spacing on x-axis labels when viewing the visualization on a mobile layout. 🎉

Oct. 1 - New Visualization: COVID-19 at Big Ten Conference Schools

The University of Illinois has been widely reported in national media for testing every single student twice a week – how are they doing? How do they compare to their peer schools within the Big Ten Conference?

The newest 91-DIVOC visualization explores COVID-19 at Big Ten Conference Schools, tracking the number of confirmed cases of COVID-19, total COVID-19 tests administered, and the test positivity.

As with all of the visualization, the visualization is updated daily. :)

Sept. 30 - Removed Filter on “Top 10” on Normalized Graphs

In the early days of the COVID era, The Holy See (Vatican City) confirmed 3 cases on March 24th, 2 more on March 28th, and another 5 cases in April. At that time, the Holy See dominated the normalized cases having an official population of just 799 residents.

As part of today’s update, I removed this filter completely – the normalized data no longer filters small countries. This is notable as Johns Hopkins University tracks the microstate of Andorra (Wikipedia) – with a ~100 new cases and a population of 77,543, Andorra has been the location of one of fastest spread of COVID-19 on a population-normalized basis (91-DIVOC Graph).

Additionally, a few tooltips were cleaned up. Look for a brand new visualization tomorrow! :)

Sept. 18 - Normalized “Top 25” shows data based on normalized-values

Selecting “Top 25” (or other similar options) on a normalized chart now works more as expected, showing the top 25 normalized values instead of the top 25 raw values.

Sept. 10 - Added Normalized/Non-normalized Statistics

For all mouseovers in area where the population is known, the mouseover will provide a normalized value for the data (ex: “3 cases /100k”) in addition to the raw data. When viewing a normalized chart, the raw data value is also given.

Additionally, the data stream from Our World In Data is now updated daily providing global testing data. 🎉

August 7 - US-Total, Computed

Johns Hopkins University has always reported the “United States” as part of their list of countries and, in a second dataset, reported the 50 states and US territories as individual locations. The number of confirmed cases and deaths between the “United States” total and the sum of the 50 states’ data has always varied slightly, accounting for the cases is the US territories, cruise ships, repatriation flights, and more. Previous cases of large discrepancies between the data have always been paired with a spike in a single state data where a number of previously unreported cases were reported in a single day.

Recently, the difference between the sums (particularly in the number of deaths) has become significant without a known cause. To help dive deep into this difference, I’ve added a new “Data” value in the graphs of the US states:

August 5

Over the past few months, one of the most common question I was asked was “how did you get started?” or “how was this created?”. Over the past week, I created a video that dives into how 91-DIVOC was created:

This is my first exploration of video, so I’d love your feedback! I believe it should give you some insight in how this visualization is created. :)

August 3

July 30

July 29

July 27

July 25

July 24 - Quality

In all of these visualizations, I created visualizations that helped me make sense of the COVID-19 data. At this point, I can spend hours diving into the data and there’s few questions that I am unable to answer using 91-DIVOC graphs. Therefore, for the next week, I’m looking to explore better documentation of this tool, fix any remaining bugs, and focus on quality improvements. As part of this, I have already:

Additionally, I’ve started some initial work on some visualizations beyond COVID-19. If you’d like to get a once-a-month update on data-forward visualizations, I have an e-mail list that I’ll be providing a monthly update of all the latest visualizations to nerd out with.

Let me know if you find any bugs and thanks for all the support! :)

July 21 - Interactive Visualization of COVID-19 in Illinois

July 20 - Dynamic Label Placement

July 18

July 17 – Our World in Data

This is the most significant update to 91-DIVOC in quite some time, adding a new data source from Our World in Data (Oxford University, et al). The Our World in Data dataset provides testing data for a number of global countries, allowing for “COVID-19 Tests” and “Test Positivity” in the “Data” selection on the countries graph, and other changes:

As part of working with Our World in Data, a few bugs were fixed and other minor display changes:

The default data selection is now “Johns Hopkins & Our World in Data”, using Our World in Data for the countries graph and Johns Hopkins for the United States data. The data source will always be displayed at the bottom of the graph.

July 16

July 15 – GIF and WebM Animation Saving

Animations can now be saved right in your web browser!

In addition, the saving of PNG images also improved:

July 13 – Improved Tooltip

July 9 - Region Enhancements

The most recent update enhances the region selections – we can explore now, for example, how the United States might compare to various WHO regions as a whole.

Additionally, the code was refactored to allow for better exploration of regions in future updates:

July 6 - Additional Global “Show” Options

July 5 - Animation Optimization

July 4 - “Add Additional Data” Improvements

Many thanks to @TheWheelMe for the initial report.

July 2 - Add Additional Data

July 1 - CSV Export

June 30

June 29 - Generate Report

June 26 - Released 91-DIVOC-04

June 23 - Added EU

June 18 - Default to Linear

June 17 - Add US States to Countries Map

June 16

June 14 - US Regions

June 11

June 10 - Multiple Data Sources

June 9

June 7 - Global and Europe

June 3

June 1

May 27 - Cumulative Mortality Rate

May 26

May 24 - Test Positivity Rate

May 22

May 21 - Saving Image

May 19 - Easy Multi-Selection on Desktop

May 18 - Animation

May 16+17 - Right-align by Date

May 12

May 11 - Weeks Matter

May 10

May 6 — Cases per Day

May 4 — Individual Visualizations

April 30 — Mobile Improvements

April 24 — Hospitalizations and COVID-19 Tests for US States

April 21

April 20

April 17

April 16 — One Week Average Cases/Deaths

April 15 — Overview Page

April 13

April 12 – Multiple Highlights

April 11 – “Highlight Only” View (and Mobile Improvements)

April 10 – Dynamic Trendlines

April 9 – Axis Zoom

March 21 – April 8

March 21 – Initial Release