Reduce waiting queue at supermarket from Corona with Python-Webapp

[Chris Angelico]

On Tue, Mar 17, 2020 at 7:54 AM Orges Leka <orges.leka at gmail.com> wrote:
>
> Am Mo., 16. März 2020 um 21:33 Uhr schrieb Chris Angelico <rosuav at gmail.com
> >:
>
> > On Tue, Mar 17, 2020 at 7:21 AM Orges Leka <orges.leka at gmail.com> wrote:
> > > For the getting enough people to use it, I think word-of-mouth should
> > work,
> > > as it would help those who use it, plus it reduces the chance of physical
> > > contact, so there is a win-win situation in using the app.
> > >
> > > Maybe if someone from the media promotes the app, this should boost it
> > also.
> >
> > In order to be usefully able to predict how many people will be at a
> > location, you'd need an appreciable proportion of them to be using
> > your app. Let's say you accept a 5% saturation (which is pretty low -
> > if only 5% of people use the app, there's still a LOT of uncertainty
> > in the estimated figures). Do you think you'll be able to get to the
> > point of having 5% of *all shoppers* in an area to start using your
> > app? That is a HUGE number of people to start using an app, and even
> > then, it would only give a low degree of confidence.
> >
> > To the early adopters, your app is close to useless. That means word
> > of mouth isn't going to be very strong. It's something that depends
> > entirely on already having lots of users.
> >
> >
> Following your reasoning, then radar detection apps / wikipedia / facebook,
> which crucially depend on user generated content should not work....
>

I don't know very much about radar detection, but I suspect that it
can be useful even if only a very small percentage of people use it.
Same with Wikipedia. (Facebook isn't useful even now, so that doesn't
really count.) But to be able to recognize when a shop is going to be
busy and when it's going to be quiet, you actually need to be able to
predict the movements of a lot of people, which means you need either
a very large proportion of people to be using your app, or some other
source of data.

As an example, consider: Suppose a mere 100 people are using
Wikipedia. Those people can create articles (perhaps lifting content
heavily from other sources, if the licenses are compatible), read
articles, edit them, etc, etc. It's not going to be a huge success,
but it is at least useful. And the 101st person to start using
Wikipedia can see some informative articles and learn from them,
and/or see an inaccuracy and correct it. It has value.

But consider that 100 people are using your app. If each person plans
to go shopping every two weeks (yes, I know a lot of people shop more
frequently than that, but ideally they'll be reducing the number of
trips for the same reason that they're trying to go when it's quiet),
then you have an average of 7-10 shoppers per day. Distribute those
shoppers among a set of time slots, and most likely you'll have
between 0 and 3 registered users in each time slot. Now consider: in
the scale of shop activity, how much can you learn by knowing the
movements of just three people? The entire point of the app is to
change people's actions, so you can't assume that the figures you have
are representative. The app actually has no value whatsoever with just
100 people using it.

In order to be useful, you FIRST have to achieve a significant level
of usage. That's extremely hard for a brand new app, so you would need
to leverage some existing data or usage somewhere.

ChrisA