README.md

# scripts for data fusion
Anaconda with python 3 is recommanded for running python scripts  
To dump the data, please use valentino and the 'export to csv' option. The order of the columns should be the same as that in the database.

runing order:  
python:  
main.py  

matlab:  
importcsv_and_7  
importcsv_ios_7  
upsample_and_8  
upsample_ios_8  
gpsExtracion_main_9  

steps before running main.py  
1, dump location data from database as location.csv  
2, dump activity data from database as activity.csv  
3, dump wifi data from database as wifi.csv  
4, dump phq9 data from database as phq9.csv  
5, dump user mapping data (depressed) from database as mapping.csv  
  

# to split big csv file into small files based on uid
'1) split_activity.py'  
'1) split_location.py'  
'1) split_wifi.py'  
input: activity.csv, location.csv, wifi.csv  
output: uid*.csv in sub-folders 
location: android/uid*.csv & ios/uid*.csv  
wifi: ios_wifiraw/uid*.csv & android_wifiraw/uid*.csv   
actiity: 

# deduplication 
'2) removedup.py'  
input: uid*.csv with duplications  
output: uid*.csv w/o duplications  
location: ios_remdup/uid*.csv & android_remdup/uid*.csv   
wifi: ios_wifiraw_remdup/uid*.csv & android_wifiraw_remdup/uid*.csv   

# look for a corresponding latitude and longitude match in location data at a given timestamp for a given BSSID in wifi data.
'3)  mapgpsdata.py'  
input: location/uid*.csv, wifi/uid*.csv  
output: matchedgps_androidcheck.csv, matchedgps_ioscheck.csv  

# merge mapping files from ios and android
'4)  mergecsv.py'  
input: matchedgps_androidcheck.csv, matchedgps_ioscheck.csv  
output: matchedgps_final.csv  

# cauculate the GPS location for BSSID
'5)  calculatemedian_latlong.py'  
input: matchedgps_final.csv  
output: bssidmappingfile.csv      

# replace the BSSID in the raw wifi file with coordinates
'6)  processwifi.py'  
input: bssidmappingfile.csv, wifi/uid*.csv    
output: wifiprocess/uid*.csv  

# data fusion - merge data from diffrernt sources
importcsv_and_7.m  
importcsv_ios_7.m  
input: wifiprocess/uid*.csv, location/uid*.csv, activity/uid*.csv
output: gps_feature_and.mat, gps_feature_ios.mat  

# data fusion - upsampling 
upsample_and_8.m  
upsample_ios_8.m  
input: gps_feature_and.mat, gps_feature_ios.mat  
output: gps_feature_and_up.mat, gps_feature_ios_up.mat  

# feature generation
gpsExtracion_main_9.m  
feature_extract.m  
input: gps_feature_and_up.mat, gps_feature_ios_up.mat  
output: dSet_and.mat, dSet_ios.mat  

# tool scripts, don't need to take care
dbscan.m  
gridLegend.m  
lldistkm.m