[tor-project] Measuring Tor's impact in the world?

[Nathan Freitas]

On Fri, Apr 29, 2016, at 01:14 PM, isabela wrote:
> So I thought of the following question: 'What happens to the tor network
> when there is an event happening in the world?'

This is really great, important work. 

> Then I thought of what world events are predicable that we could check
> for Tor's impact? elections...
...
> Of course we should also take a look for the same thing when
> unpredictable events happens.. like the Egyptian revolution or a
> shutdown of services for whatever reason (whatsapp in Brazil).

I think that shutdowns of service of blocking of specific apps are
another important metric. This is often when we have seen growth in
Orbot downloads in a specific region.

On that note, I am happy to share more detailed Orbot download
information with you as part of this effort, to see how that correlate's
with general Tor usage in a specific region.
 
> Why I am sharing this? Because there are tons of smart people in this
> list and maybe someone has a different way to measure tor's impact in
> the world.

I think beyond a list, there a great deal of researches at places like
the Citizen Lab and the Berkman Center that would be eager and very
capable to help with such things, as well as separate noise from actual
events.

> 
> Or maybe someone knows how I could automate such a thing :) (api
> somewhere?)

Unfortunately, I don't think this can be automated, but at the same
time, I think just showing a few important, meaningful correlations per
year would have a big impact.

> Or maybe ideas of other points of data for reference. For instance, I am
> now thinking if x users connected on Tor means a lot or a little in a
> country.. that will depend on how many people are online, the number of
> internet penetration for that country.

That was definitely running through my mind as I was looking at the
graphics. 200 users may seem small, but depending upon Internet usage,
desktop vs mobile, and whether you have the right 200 people are all
important context to have.

> Yes, I know this is very simplistic way of talking about something as
> big as 'impact in the world'. But I have seem this information being
> useful in other scenario, and it did helped a lot to pass the message of
> what value people should think of when thinking of the product.

This is exactly the way my wife who works on multi-million dollar public
health projects measures their impact of programs to get people to quit
smoking or to reduce infection of certain disease. At some point, it all
comes down to "did less people die in X region from this terrible
thing?", or "how much money did we spend per person who did not die?".
If that metric is inline with what the funder expected, then you get
more money, and less people will die.

> talk about it. And for a while the data science team did only this,
> publish correlation data of what happened at twitter during global
> events. I believe it did worked because that information was what people
> (media, mouth to mouth etc) start to pick up when anyone would talk
> about twitter.

I completely agree, and am happy to support the effort to have the same
kind of impact with Tor, Orbot and all related onion efforts.

+n