Show HN: Covid-19 Treatment Dashboard | Hacker News Reader

Show HN: Covid-19 Treatment Dashboard

quiverquant.com

137 pointsgreatwave141 comments

Making this information accessible is amazing work, thank you!

But, alas, it loses a lot in the UX department:

1) It's low-contrast. This makes it hard to read for a lot of people, unnecessarily so.

2) It uses a font that is hard to read. Not sure if it's font rendering, or if the font is just bad, but... just a plain Sans Serif would be fine?

3) The gradient for the map is pretty much not fulfilling any function, because you can't read it with any kind of precision. You'd be better off with 4 distinct gradients. Or better, separate colors.

4) If I can nitpick: The title has 3 different font faces, with misaligned baselines. Please don't do that :)

Ironically, the scale legend in the chart actually uses a font that's very easy to read

I wonder if the thinking was to make it deliberately uninviting and difficult like the Bloomberg terminal to make people feel cooler while using it (which is a legitimate thing - i'm not saying that glibly at all. People that master the bloomberg terminal feel like wizards which certainly contributes to lockin, and that sort of needlessly difficult UX is somehow associated with being "cool" or being like a "real command center" somehow)

While I agree with the GP, I don't agree with the jab at Bloomberg terminal. There is nothing needless in Bloomberg terminal - it comes with a special keyboard and everything is keyboard driven. There are no animations such as the ones you see in macOS and Windows that take up 300ms. There are no distracting popups and notifications. It is probably the most incredible feat of UI/UX design...ever, perhaps airplane cockpits and nuclear power plant control rooms would rival.

Just because UI is dense does not mean it is not easy to learn. Dense UI provides more information than something with 5 levels of hierarchical layouts, each screen animating with 300ms of bullshit animations. Furthermore, learning curve should be acceptable for people that truly want to use a tool on a daily basis (for e.g. emacs or vim).

There is a reason why people who use Bloomberg terminal will never want anything else. Here is a video of the thought process that goes into designing just the keyboard: https://www.youtube.com/watch?v=_juj1MIRJVE

Have you used Bloomberg terminal or talked to a user? Regarding lock-in: I think its the opposite - when the users don't want to switch to anything else, that's the best a company can do. That's the kind of "lock-in" every company should aspire to.

I've used a Bloomberg terminal and most of my social circle has as well. I guess maybe difficult isn't the right word to describe it, but definitely uninviting on purpose. And a lot of my friends get a kick out of how uninviting it looks and how much they feel like a wizard using it. Theres definitely an aspect to the way it looks unrefined that make people love it, and I guess a better way of saying it might be that trying to emulate that by making something actually inaccessible and unrefined is not going to be successful

I think you're attributing ugliness "on purpose" incorrectly. No one at Bloomberg exclaims "Oh gee, let's make this more ugly, people will get a kick out of it."

It's just that they haven't bloated this app with CSS styles and other unnecessary aesthetics. It works. That's all.

Thanks for the feedback. Fixed the fonts and title, I'll spend some time tonight working on the other points you brought up.

Awesome, thank you!

(You still might want to align Quiver Quantitative's baseline with the rest. It's the smallest nit of them all, but it's jarring ;)

Very genuinely curious, do most people consider these "UX" instead of "UI"?

For me, I'd like to be able to click or at least copy/paste the titles of the studies shown on the map, e.g. when I hover over houston, I can visually read the title of the study, but I can't move my mouse to click or highlight the text, or else the box disappears.

I'm very glad someone is making a dashboard that actually brings "new" information that's presented differently than other people have presented it. Lots of fairly derivative coronavirus dashboards have popped up on ShowHN over the past month or so, but this is the first along these lines.

It's probably more UI than UX, but... they're really mutually reinforcing. I'll start calling it U* ;)

FDA: emergency use authorization for hydroxychloroquine and chloroquine

https://www.fda.gov/media/136534/download

Yep, and this possibility has become a right cluster-fuck for many people that have other needs for the drug.

Many Lupus sufferers can no longer get this because of possible use (https://www.kqed.org/science/1960404/at-kaiser-trumps-pharma...)

I know people that have this condition and have been told there's none available at pharmacies when in fact it's in stock (based on how they resolved their problem).

And how did they resolve their problem?

They asked until they reached the right person? Its not that uncommon of a pattern, sadly.

Basically this.

Here’s a bit on that French doctor who started this particular craze: https://slate.com/news-and-politics/2020/03/didier-raoult-hy...

He doesn’t believe in evolution, nor climate change. At the end of January, he called warnings of a pandemic „crazy“. He has published 3000 papers in the last ten years. And even the photo of him is quite funny if you followed the news in 2016.

Judging by the list of trials, an awful lot of resources is spent on this. I wonder what the opportunity costs are.

There's a ridiculous controversy in France about this guy. Lots of French people see him as the misunderstood genius that could save us from the virus, if it weren't for our evil government (obviously sold to the pharmaceutical companies). We even have celebrities (actors, rappers, humorists...) begging the president to listen to this doctor. We reached the point where we trust "celebrities" more than the scientists leading the effort to fight the pandemics.

I watched the Greek daily COVID-19 update, and the epidemiologist in charge was asked about chloroquine specifically. He said that we are administering it to patients, along with 1-2 other drugs (if they want it), but the results haven't been conclusive yet.

We "only" have 2-5 deaths a day, so it's not a huge sample size, but if the drug were extremely effective, you wouldn't need a very large sample size to see a difference.

> We "only" have 2-5 deaths a day, so it's not a huge sample size, but if the drug were extremely effective, you wouldn't need a very large sample size to see a difference.

I'm not so sure about that. If the goal is to get from 97% case survival rates to 99+%, the results may not be all that obvious from small sample sizes.

Minor bugbear: the thin font used does not lend itself to rendering nicely in Firefox 74 running on a GNOME 3.36 linux install with subpixel font antialiasing: https://i.imgur.com/LauxkrQ.jpg

I made it more readable it for myself by switching to grayscale antialiasing, but it's still not a very readable font as it's quite small and thin and is light-ish blue against dark-ish blue background - standard sans serif, slightly larger and with more contrast in colours probably better for accessibility

Even when it renders properly it's still not very easy to read.

Thanks for the heads-up. Changed the table fonts to Verdana, hopefully they're easier to read now.

I noticed that this study is not on the list:

https://clinicaltrials.gov/ct2/show/NCT04264533

What is the criteria for choosing studies that show on this list?

Also the University of Minnesota Losartan trial: https://clinicaltrials.gov/ct2/show/NCT04312009

Probably don't have to look far to find another dozen or so.

First thing I checked, sadly missing considering the amount of medical field data on it coming out of China/East.

Hey, this looks really good! Didn't even know that the information about those studies was accesible. I don't understand the phases meaning, can you explain them? Maybe a table or tooltip somewhere could help.

Thank you very much! That’s a good idea... I will think about the best way to incorporate that information on this dashboard, but for now you can find it at https://www.nccn.org/patients/resources/clinical_trials/phas...

Thanks for the site.

Bug report: Looks like sort by expected finish works on strings instead of dates.

Sorting by expected finish date seems be broken - it sorts by strings

came here to say this

Looks cool. There's a lot of double entries. Is that in the data or is it a glitch?

The table "News articles on COVID-19 Treatments" shows a duplicate "FDA authorizes chloroquine for use in treating coronavirus"

This is great, is there any information somewhere about the efficacy of each treatment? I'm on mobile and may have missed it.

All I see is a login page?

Minor feedback - I wish I could click on a trial on a map to go to that trial page.

Great job you should make the lists scrollable but it looks great !

As per www.worldometers.info, Russia seem to have relatively low death rate (0.06 per million of population), despite having a border with China, no early distancing measures, public transportation in the cities, etc. Could be just accident, but they did one interesting thing - recommended a nation-wide treatment strategy early on.

The confirmed cases number is effectively useless without models for extrapolating into an estimated infection count. Presumptive cases aren't bothering to be tested, asymptomatic cases rarely tested, lack or abundance of supplies differ by country and region, etc... We have very little idea of the prevalence in any given country - confirmed case count is just a floor.

If you notice, I'm talking about number of deaths, not the number of cases. I do understand the limitations of confirmed cases stats.

Russia shut its border with China super early, which helped. However, it better be careful now now that its population is unprepared mentally for a potential outbreak seeded by European travelers weeks ago.

As far as I understand the mentalitet there, you don't have to worry too much.

Reply on news.ycombinator.com