Mobile-Phone Swarm-Motion Detection, Mass-Data-Analysis and Visualization
WTF?
Imagine you have some kind of Motion-Data. The motion from Mobile-Phone-Devices over an geographic Aera.
We would need hundrets or thousends of such kind of ESP8266-Mobile-Device-Scanners to cover a big Town ( like Munich ) .
Finally we took the Data from some of our scans..and generated the rest of this kind of data. This way the data we observe here is a simulation ...some kind of fake-Data. So we take this data as given!
But bear in Mind - the technology behind all this is REAL! The basic Concept on how to Measure such kind of Device-Location is described here: https://github.com/iCounterBOX/Trilateration-with-NodeMCU-8266
The Experiment
Our assumption is to have hundrets of those Scan-Boxes over an area..e.g Munich:
and finally we want vizualize how a target-Group ( e.g. devices at 6:00 AM ) are MOVING across this area! An animated simulation...
Play-Data
each Scanner-Device 
is getting a random-ID over an Geo-Location Area. In this experiment distributed over Munich. As each Mobile device has a unique MAC - a Grid of such ESP-Scanners is easily able to store this in form of "Trails". e.g 100 Devices detected in sector A ( 6:00 AM ) ... 1 Hr later those 100 Devices also detected in Sector B - this way we have a MOVE vom Sector A to Sector B...then form Sector B to Sector C...etc etc
src,dst,time_src,time_dst,day,amount
12591,8175,6,6,MON,4
8175,15677,6,7,MON,11
15677,15646,7,7,Mon,15
Each Scanner will have this randomNumber-ID AND a fixed GeoLocation ( e.g. positioned in a Street-Lamp ! )
Our goal is to understand and to visualize when, where and in which amount our swarm is moving. Will it be something like this? A Move from Sector-Scan to Sector-Scan. Detecting where people live, stay ( how long?), leave, work, celebrate,....
considering all this we generated some random ( fake ) data of such kind of "Trails". with trailNr,src_dst,src,lon,lat ( from the scanners)..and some Timestamp-info when the scan was done
trailNr,src_dst,src,lon,lat,date,day,hour,minute,count,v,s,t
1,8174_8185,8174,11.410571516906163,48.15152679449321,2017-05-01 06:00,MON,6,0,22,0,1999.0,60
1,8174_8185,8174,11.415943928198173,48.15145741662311,2017-05-01 06:05,MON,6,5,22,0,1999.0,60
1,8174_8185,8174,11.421316339490183,48.151388038753005,2017-05-01 06:10,MON,6,10,22,0,1999.0,60
1,8174_8185,8174,11.42668875078219,48.15131866088291,2017-05-01 06:15,MON,6,15,22,0,1999.0,60
1,8174_8185,8174,11.4320611620742,48.1512492830128,2017-05-01 06:20,MON,6,20,22,0,1999.0,60
1,8174_8185,8174,11.43743357336621,48.151179905142705,2017-05-01 06:25,MON,6,25,22,0,1999.0,60
1,8174_8185,8185,11.43743357336621,48.151179905142705,2017-05-01 06:29,MON,6,29,22,0,1999.0,60
1,8185_15638,8185,11.43743357336621,48.151179905142705,2017-05-01 06:30,MON,6,30,20,0,1581.0,60
1,8185_15638,8185,11.441436997212305,48.15022748181894,2017-05-01 06:35,MON,6,35,20,0,1581.0,60
1,8185_15638,8185,11.4454404210584,48.149275058495164,2017-05-01 06:40,MON,6,40,20,0,1581.0,60
1,8185_15638,8185,11.449443844904495,48.1483226351714,2017-05-01 06:45,MON,6,45,20,0,1581.0,60
1,8185_15638,8185,11.45344726875059,48.14737021184762,2017-05-01 06:50,MON,6,50,20,0,1581.0,60
1,8185_15638,8185,11.457450692596685,48.146417788523856,2017-05-01 06:55,MON,6,55,20,0,1581.0,60
1,8185_15638,15638,11.457450692596685,48.146417788523856,2017-05-01 06:59,MON,6,59,20,0,1581.0,60
Assuming that our ESP-Scanner-Devices are distributed over the town - we will have up to 60m between our scanner devices ( or more if we can use far-range-transmitters). To visualize this in Folium we extra generate some random-Points between those Scanners.
In Reality: as much of those Scanners would be distributed - as better and precize the Scan-Results will be. One scenario would be to add this devices into Street-Lamps:
pic is showing a prototype we made some year ago. the followUp model was better integrated.:
only the Antenna was on Top of the Lamp. (This all is Experimental Stuff - just to elaborate "what is possible?")
Visualization
There are planty tools and stuff on how to visualize Swarm-Movements..
QGIS ( https://www.qgis.org/de/site/ )
py/folium ( https://python-visualization.github.io/folium/ )
https://github.com/keplergl <--- this will definitely my next tool to visualize those trails. It needs a bit an other format of the src/dst etc.
So for the moment this experiment here base on PYTHON, QGIS and Folium!
From random-Trail to MAP