Crowdsourced data collection: what to get?

Given the epidemiological facts around the virus, what type of data would be most useful to collect in a crowdsourced way in order to help address the crisis the “best” way? “Best” is left undefined as there will definitely be trade-offs. One might prioritize minimization of deaths, health of doctors, getting schools back on, companies back online, etc.

Please respond below with some (not-necessarily prioritized) ideas.

  • composition of households (age distribution in particular)
  • coughs and other symptoms
  • temperature
  • professional activities

Maybe gather the name of the local hospital so that could be used in demand/ supply/ notifications?

I would also recommend that we gather each individuals most relevant Medical ID. So in UK for example, each person has an NHS ID as below. That will be useful in many scenarios when we need to join/ share bottom up data with the top down equivalent.

Consent to making their data public if tested positive
Consent to making their data public if dead (if needed)

1 Like

Agreed. As a placeholder down the track, there will come a time when financial transactions data will come into the mix. There will be many aspects to this; but initially and simplistically if I make a bank transaction at a supermarket that data can be used to help in my assertion that I was there. I’m flagging this mainly for downstream needs; there are two ways my bank data can be accessed; one with my consent (via Open Banking API’s in UK for example); the other getting it from the bank. The former is most definitely the MyData preferred architecture.

Swiss flu tracking project

Are you willing to go out? (combined with “I think I had it”, this might mean virus-resilient people are willing to go around town more to help others - after testing of course)

How many people are you self-isolating with? what are their ages?
Is this a "perfect seal? (maybe you are interacting with another household)

have you lost your sense of smell?

  • Parental consent if <16
  • Consent period (unlimited vs fixed period)
  • Right to withdraw consent

Antwerp is a world-renowned center for infectious diseases. They are crowdsourcing this data:

social contacts and mobility:
who visited me/provided me something, where
who I visited/provided something, where
when and where I bought groceries
when, from-to where I took a bus, a taxi, Uber
when, where I took a walk alone, with my dog

are mobile phones logs useful?

can’t access this URL, is it me or wrong URL?

This one: < bit late for implementation but this tool try to follow this reasoning

A bit hard to catch up, if the purpose is great, we have to keep in mind that these data could be gold for some companies and are a heavy risk for the people - imagine a social security number is used to impersonate you, or falsly identify someone to accuse this person of contaminating people, that could lead to a lot of damages.

  • which data are really needed for the purpose?
  • how to avoid direct / easy identification of the persons? (no use of medical ID or security social number)
  • how to protect the database (segregation of data, security measures)

A possibility is to gather anonymized data for the researchs (age range, heart rate, temperature, with how many people do I live…), and that personal data are stored in local by the concerned person (like identification, exact age, GPS tracking, identity of people I met), which could be useful in specific cases (inform quickly potential exposed persons of a contamination).

yes, definitely. this was a thread where precisely we collect what might be useful, and then we can assess real need/risks/proportionality/minimization etc

and welcome!