jayrbolton / covid19-growth-dashboard Goto Github PK

View Code? Open in Web Editor NEW

4.0 4.0 1.0 3.19 MB

COVID-19 tracking dashboard

Home Page: https://www.covid19tracking.org

License: MIT License

HTML 0.31% TypeScript 94.00% CSS 5.69%

coronavirus coronavirus-tracking covid-19 covid-tracking covid19 preact tachyons typescript

covid19-growth-dashboard's Introduction

Hello:grey_exclamation:

I am a software engineer living in Oakland, CA with interests in computational biology, creative coding, and sharing knowledge.

I currently work on infrastructure for genomics data analysis at Karius Diagnostics.

I've been spending some free time solving Rosalind bioinformatics problems and writing a small puzzle game in Unity involving cats.

Lately I've enjoyed programming in Rust and C!

Other projects:

Polygram for interactively generating mathematical animations
Monkeypatch CMS for live editing web pages
A COVID-19 Dashboard for tracking US and worldwide growth
1d Cellular Automata simulator in javascript
Dontdie - a sass startup runway calculator view
Linear partitioning algorithm in javascript
Hadrosaur is a Python library for managing large amounts of computed files
Brontosaurus is a Python library for creating JSON RPC APIs
Smooth scrolling library in javascript
browserify-crawl and postcss-walk
ordered-hash data structure for javascript
Some flyd (functional reactive programming) libraries such as flyd-crud, flyd-undo, flyd-ajax, and others
Langton's Ant in Python
I worked for 6 years on a startup called CommitChange which lives on as OSS called Houdini
I helped create RallyStarter
I did a little work with Dat Project, a very cool P2P framework

Some incomplete projects that I may return to:

dat-wot is a WIP p2p cryptographic communication library for Node
Ukkonen's Algorithm in Node

covid19-growth-dashboard's People

Contributors

Stargazers

Watchers

Forkers

briehl

covid19-growth-dashboard's Issues

Compute aggregated stats

For the current search, show aggregated stats at the top for all results below.

If no filter is applied, you see worldwide stats. If you filter only US, then you see US-wide stats.

More frequent updates

We should try to use the JHU servers for data sources, which get updated more frequently.

Can always fall back to the github repo urls if the request fails

Add active cases and convert confirmed->active

In normalize-data.ts, compute a totals.active array to be (confirmed - recovered - deaths) for every day in the time series. Also add corresponding metrics to row.averages, row.maxes, row.percentages, etc

Change all cases of "confirmed" in the UI to the "active" numbers.

Draw bar time-series chart

Draw a simple as possible time-series bar chart of total cases per day for each region

Fix localstorage caching

At some point this started messing things up for me. I removed it completely for now.

Caching is always 🤦‍♂️

Paginate

Show the first 100 results or so. We don't need actual pagination. Just a "show more" button or similar is good.

Better visual grouping of "average new cases"

Fix average new cases

The average new cases over all days and over the last 7 days should be calculated on the differences of the active numbers, not the confirmed numbers. This is also used in the sort by "recent growth" function.

Mobile compatibility and responsiveness

Make sure everything looks good on mobile. We can do this later on when more features are in place.

Tachyons will help

Disclaimers, attributions, sources

https://medium.com/nightingale/ten-considerations-before-you-create-another-chart-about-covid-19-27d3bd691be8
https://blog.datawrapper.de/coronaviruscharts/

Compute average new cases

Over all time and over last 7 days. Display these stats in the RegionStats component

normalize-data.ts could do the computation upfront

Better visual grouping of "average new cases

Add percentage to active row.

In the totals list on the left, active doesn't have a percentage. Oops!

Filters

Filter by:

Country name
Province/state name

Later on, but we should verify if these are useful:

Filter by continent or larger regions (eg. Europe, North America)
Filter by min/max recovery/deaths percentage
Filter by min/max confirmed/recovery/deaths/averages

Recovered/death percentages?

Does the total confirmed number in the data source include deaths and recovered?
Should the percentage recovered be (recovered * 100 / confirmed) or should it be (recovered * 100 / (confirmed + recovered + deaths))?

Add tooltip with values to bar chart

Mousing over a bar chart could show the values in that bar. Just something like:
total: x
active: y
recovered: z
deaths: a

Should have a translucent background to not cover the chart, or always be positioned away from that bar (maybe off to the side and an arrow pointing at it?)

For averages, show exact number of days, not "all time"

All time is not accurate

Upcoming breaking changes from JHU

CSSEGISandData/COVID-19#1250

Files will be renamed with columns moved around. Most importantly, recovered numbers are going away

If we want to keep this dashboard going, I say we move away from JHU as the primary data source and maybe switch over to covidtracking.com for the primary UI. We can still keep a secondary dashboard of JHU numbers that will probably be more unreliable

Make more colorblind friendly, and avoid use of red-adjacent colors

Some colorblindness guidelines here: https://www.toptal.com/designers/colorfilter/

Avoid the red + black combo (technically it is orange, but close enough to red to warrant criticism)

Migrate away from JHU data to covidtracking.com data

Testing numbers are very important when taking into account active case numbers.

Requires more planning and design

Data sources:

US - machine readable - https://covidtracking.com/
World - maybe not machine readable - https://ourworldindata.org/coronavirus-testing-data

Sorts

Sort by:

Total confirmed
Highest recovery percentage
Highest deaths percentage
Highest average new cases per day

There are other possibilties, but these are the useful ones that I can think of.