Wigle Wifi for Android: Suggestions

[bobzilla]

There should be the possibility to search networks nearest to the current location and/or the possibility to select only OPEN networks. So wherever I am, I know where is the nearest open network around me.

This is a wardriving app, which is collecting data. Using a network without permission is likely illegal depending on where you live.

- the upload to wigle site is buggy. It does not retain the FREE/WPA status, so, once uploaded, this information is lost. Very annoying.

Not sure what you mean by this, crypto data is logged from all submissions.

[sanna030]

There should be the possibility to search networks nearest to the current location and/or the possibility to select only OPEN networks. So wherever I am, I know where is the nearest open network around me.

This is a wardriving app, which is collecting data. Using a network without permission is likely illegal depending on where you live.

Thanks for the reply, I am not talking about connecting to any network here, I am just suggesting some hints to improve the application. and this definitely would improve it without impacting on any law of any state. *IF* and only *IF* you use in the way you suggested the information the application is collecting for you, *THEN* you are doing (maybe) something illegal (which is actualy something I could eaasily do with or without the improvement I am suggesting here, anyway).

- the upload to wigle site is buggy. It does not retain the FREE/WPA status, so, once uploaded, this information is lost. Very annoying.

Not sure what you mean by this, crypto data is logged from all submissions.

Unfortunately, this is not true at least from me. The operations I did are

1) going around collect a few WiFi network with my htc sensation
2) stop the scan once I am back at home
3) click "Upload on wigle.net" button in my app on my phone
4) the app tells me he is preparing some files and that it has uploaded them succcesfully
5) go to website, login with my username
6) use the search form and search for network discovered by me. The list returned contains all the uploaded nets, but has *NOT* crypto information correct. Both free, wpa, wep networks are all labelled WPA.

S.O. is ICS and app is updated to latest. If you want more infor to track the bug, please let me know, but it would be just enough to search the database using my username and go through upload logs, if you are an administrator.

Anyway, again compliments for the great app.

[ccie4526]

Note: user 'anonymous' is the most popular, won't be getting rid of it

Boooooooooooooooooooooo.

[pejacoby]

Note: user 'anonymous' is the most popular, won't be getting rid of it

Boooooooooooooooooooooo.

Maybe you could triangulate a location and split 'anonymous' up into 'anonymousSTATE' or 'anonymousISOCountryCode''?

[mtbrandao]

Why get rid of anonymous?
Or try to triangulate anonymous? That breaks the whole reason of not entering information and being anonymous.

[Vchat20]

I don't know if this suggestion has been made already, but it'd be cool if there was an option to tell the software to scan on a singular channel or selection of channels (given the OS allows control of this level).

(Like many people) I have a regular routine route to and from work. 10 miles each way evenly spread between freeway and in-town driving. So far after finding out about the software I have been running it nearly every day round trip and it has been a couple weeks since installing it. Yet, even today, I am still getting reports by the app of new APs being discovered. I do understand the given complexities of wardriving concerning using a single wifi device and trying to scan 11+ channels at decent road speeds. The way my OCD on these things is going, it'd be nice to be able to hit up one channel at a time for a few days, wait til the 'new' count drops to 0 consistently, and then move to a new channel.

[daveemtb]

Not sure whether this is a real bug or not, but the new/old WiFi doesn't seem correct - when I visit an area that the webmap shows has not been mapped at all before, and I certainly haven't, quite a high percentage of APs are detected as not new. Any ideas?

[logplan]

Don't know if this thread is in use anymore... but
Love the program! The battery watch that turns it off automatically is a great feature. It looks to me that my data is not saved before the program shuts down, though. Is that correct? In case, it would be nice to have it do a save before automatic shutdown.
-Bob

[pejacoby]

What would it take to support the Kindle Fire HD wifi model? My kids won one, and I was disappointed to see there are no stumbler apps for it. I've only had it for a day, so I am still looking at GPS connection and things that might make it impractical for stumbling.

[daveemtb]

It seems to me that WigleWifi for Android is logging the location of APs when first detected during a run, and not at subsequent consecutive points detected, including the point at which the signal strength is strongest.

This is reducing the accuracy of the data produced, because in dead-end/no-through-roads (which is virtually all residential roads where I live) houses are only ever approached from one direction. Thus logging more points doesn't improve the accuracy. You can observe that many dead-end roads show no APs on the web map at the "closed" end of the road. They are all bunched further up. A possible workaround is to upload, exit WigleWifi and reload when at the end of such roads, to place a new point with higher signal strength there. This can take ages in the middle of a long run, however.

Can I suggest that Wigle outputs the location of APs in the upload files at the location where the strongest signal was seen, not at the point it was first seen?

So, if AP "A" is seen at location 1 with signal strenth -95dB, then Location 2 with -85dB, then Location 3 at -75dB, that location 2 is uploaded as the location, not location 1? Even better, upload all three locations with signal strengths, for triangulation from the various points from the same run.

More information on how WigleWifi handles this kind of stuff on the Wiki would be nice.

[uhtu]

More information on how WigleWifi handles this kind of stuff on the Wiki would be nice.

https://github.com/wiglenet/wigle-wifi-wardriving

-uhtu

[vos093]

It seems to me that WigleWifi for Android is logging the location of APs when first detected during a run, and not at subsequent consecutive points detected, including the point at which the signal strength is strongest.

I think daveemtb is right. My own AP is already scanned 100+ times. One would think this makes the position very accurate. This is not true. The position is very wrong because of the "first detected" design issue. It decreases the value of all scanning data already done by the Android app. I think this is a major problem.

[MDofPerth]

Server-side uses all observations from all observers. Client side is last-one-wins currently, we can probably make that better.

I support the original poster's suggestion for triangulating, and/or using the best signal strength reading for the 'position' of a given wifi node. At present I can drive down the highway and pickup access points from a block or even two on either side of the highway. If I later drive down those back streets and pick up a stronger signal from the points located on those streets, the map still shows the Access Point as being on the highway.

My own business access point appears about a block and a half from the actual location, probably because that is where someone (not me) first picked it up.

What is the current treatment of multiple observations, and what plans are there for the software to treat observations of different signal strenght? Does it differ inside USA vs outside the US?

Cheers, MDofPerth (Australia)

[bobzilla]

Server side handles multiple observers, signal strength, a lot of hard-coded and statistical filtering, which is more than we can ask on a mobile platform which doesn't have all the data. But like we said, we can probably do something with the local data, problem is being able to grab that data and update on the fly, we're already pushing sd cards/internal memory pretty hard. Selecting out 100's-1000's of observations for a network, running through them to compute a more-likely location, will burn disk i/o, cpu, and battery. Not straightforward, but we can probably come up with something that doesn't cost too much.

[Vchat20]

Would it be possible to have something PC-side to do the extra computations needed? Given this application is more an 'experienced user' app than a 'joe user' app, should be more suitable in this scenario. I know, personally, I would not mind at all just copying a file of raw data (containing all observations and not just a single one) from my phone to my computer, run an application there that computates and finds the most likely location and uploads it, or simply does some compression on the existing data (depending on the size) and uploads then. Or something along those lines. Disk usage isn't much of a concern these days even on phone SD storage, but CPU and IO time is (as you noted). If that part can be offloaded somewhere else on the tradeoff of having more accurate reporting from the Android app, I'm all for it.