Ideas and Validation March 9, 2020

Do we need a Covid-19 tracker?


Hi Indie Hackers,

In the US right now, there isn't nearly enough tests for all those who may have the Virus.

A simple tracker might be helpful so that people can know if their area is effected, to help healthcare workers prioritize care, or to help governments make decisions based on data.

I'm thinking a public website that gives the world basic stats, and a database (perhaps just Airtable) that we can share with officials with more private information.

Has anything like this been done already? Would you help this effort (marketing/engineering)?

  1. 4

    Isn’t this already served by public health officials and researchers in most countries?

    1. 1

      No. We’d be tracking requests not confirmed cases.

  2. 1

    The data is there. You mentioned the US. Unfortunately, there is a concerted effort from Trump and his allies to bury this information. I'd say pick a state first, like California. Get a hold of either Kaiser Permanente and Blue Shield. They'll have the most data. Get it in CSV format, as this is most likely how they export their data. You would need to do this for each hospital group, and insurer for each state. Just calling 300 places and getting their data would take about a week. (I've done this for beer sales - not joking.) And you'd have to call again after that week to get next week's data.

    The requests are already being tracked on a hospital basis. (They are the ones ordering the kits).

    So, basically to start you need 1 person calling full-time, and then on-boarding them to a system that will ingest the data. Then you have to able to serve the data on a stack at scale.

    The great thing about this is that when you are done, you will have a monopoly on request data. The sucky thing about it (having done this for an app called MyChamberApp) is that you will be surprised at how many "no's" you get, and so your data set won't be complete. I'd say in a best case scenario, you'll get 10% of the data you need. The other 90% of the data can be extrapolated using a stats model.

    Good luck!

  3. 1

    Now there are so many Coronavirus Trackers, would be very difficult to get traction if you start now. I made one as for Visa List users it very important and already have 400K MAU.

    1. 2

      This is different. We’d be tracking requests not confirmed cases.

      1. 1

        Oh I see.

  4. 1

    There's already a ton of them out there since it's been going on since January in a lot of the world. Trackers tend to rely on public data, so things are heavily dependent on whether the government is actually publishing data in the US.

    A few examples:

    1. 1

      This is different. We’d track the requests for tests, not the positive results.

      1. 1

        Would be a lot of false positives. Why would I need to put my request there?
        The officials will be too busy, btw.

        1. 1

          Officials won’t be busy. They’ll be swamped. It’s much easier to filter and scan a spreadsheet then it is to field thousands of phone calls.

          Yes, there will be many false positives, but the data would be helpful regardless. Much easier to find trends with data and filter for the false positives than to go on no data.

          1. 1

            Well, the swamping is actually solved, I believe, with the existing system. I have no idea what country you are talking about (is it the only one?), but here in Italy, everyone should notify the family doctor first. The doctor is one to contact the next level of the system.

            In many countries, almost everything is heavily regulated. In the US, I read in the NYT, the officials ignored the alarms from one laboratory just because there was no appropriate protocol for that.

            I'm just expressing my doubts. You can ignore them, too :))

            1. 1

              Hi Dimmi,

              Thanks for the response from Italy. How are you coping?

              I’m in the US, the virus isn’t being taken seriously enough yet. Testing is really really hard to get. We have a decentralized (and pretty bad) healthcare system, and we’re not ready for this.

              So, yeah, this isn’t “the solution”, but it may “a solution”, or at least it may be helpful. We’d need to make it clear on the website that users should not rely on filling out the form, and contact a doctor as well, but a simple spreadsheet might help us get a handle on how fast the outbreak is spreading.

              What do you think?

              1. 1

                Hi Josh,

                I have to say I thought in the beginning that Italians took the things too seriously, there was no reason to close all schools, universities, bars, restaurants, all public spaces at once for one month, killing the economy. The main reason is that the officials reported the average age of the deceased ones is 81. And I'm still not sure if the response to the virus is justified.

                But the thing wich impresses me is the self-responsibility of everyone, starting from regular people. And the motto: "Let's stay distant today to hug each other stronger tomorrow".

                Back to the topic. Say nothing can stop you from this project. I'm not sure a sic person with a difficulty to even breath can think of self-reporting. But a healthy neighbor or a volunteer can do the job. Your task would be to use a probabilistic model to approximate reported numbers to the real ones (taking into account redundancy, population, etc).

                Good luck!

    1. 1

      This is different. It’s about tracking requests for tests.

  5. 1

    I'd like a qualified expert in virology, not tied to any particular government, assessing the risk and sensible precautions to take in each country based on facts.

    I don't think a map which I have to analyze myself as a layman is helpful and could be harmful as it might create panic when there doesn't need to be, or complacency when there should be care taken.

  6. 1

    How would this tracker work exactly? Hard to say if it's been done before without details about how it'd function. Also, here's a guess at some of the main challenges:

    • Growth. As always, how do you get lots of people to adopt this?
    • Utility. How, specifically, will this help people? IMO should attempt to explain what people in what roles are going to be able to make what decisions as a result of this.
    • Disutility. Are you sure this won't cause unnecessary fear or panic as the virus continues its inevitable spread to every town and city in the country?

    There's also the issue of accuracy if this is all self-reported data.

  7. 1

    There are already some of those about

  8. 1

    It might be worth looking to see what @kbasset has done with

    1. 1

      That’s useful, but not the same @rosiesherry. We intend to crowdsource “who needs a test”, not “who’s been tested.” In the US, it seems likely that under 1% of those who have it were tested. This might help.