Mobile-Phone-Based Remote Patient’s Vital Signs Monitoring and Automated Alerts
What is mobile phone data? - ESCAP
-
Upload
khangminh22 -
Category
Documents
-
view
3 -
download
0
Transcript of What is mobile phone data? - ESCAP
What is mobile phone data? Overview of data generated by mobile communication technologies
Siim EskoPositium
Estonia
@positiumwww.positium.com
10/06/2019 Mobile Phone Data Introduction
If in doubt about MPD
Consult the UN GWG Handbook on Mobile Phone Datahttps://unstats.un.org/bigdata/taskteams/mobilephone/Handbook%20on%20Mobile%20Phone%20Data%20for%20official%20statistics%20-%20Draft%20Nov%202017.pdfor among the course materials
10/06/2019 Mobile Phone Data 2
Mobile Positioning Data (MPD)
• Active positioning data• Mobile Positioning System• App based positioning
• Passive positioning data• CDR’s, IPDR’s• Location area update, signalling data
10/06/2019 Mobile Phone Data Introduction 3
Types of Data• Call Detail Records (CDR)
• Outgoing• Calls, SMS
• Incoming• Calls, SMS
• Internet Protocol Detail Records (IPDR)• Mobile internet usage
• Location area update• LA change due to movement• No activity > 2hr => LA update
• Signalling, Abis, Probe data• Virtually any activity that involves communication between the device and the
network• Signalling includes all types above + additional – handovers, phone switch on, etc.
10/06/2019 Mobile Phone Data Introduction 6
Data Events• Mobile-Originating Call start• Mobile-Terminating Call start• Mobile-Originating Call end• Mobile-Terminating Call end• Mobile-Originating SMS• Mobile-Terminating SMS• IPDR initiation• IPDR termination• Location area update• Location area change• Handover event• Device turn-on• Other
Data types include different events
10/06/2019 Mobile Phone Data Introduction 7
CDR
IPDR
LU
SIGN
ALLIN
G
Main 3 Attributes in Raw Data
ID Timestamp Cell_id MCC
IMSI based(pseudonymised) ID of the subscriber
Time of the event Location of the event• Cell_id reference• Geographical
coordinates• Geographical
coverage area
Mobile Country CodeNecessary to distinguish domestic from inbound roaming
10/06/2019 Mobile Phone Data Introduction 8
Availability of Different Types of Data
Data type Availability in MNO Approx. Avg number of records per subscriber per day
CDR outgoing Always stored (billing) 2-3CDR incoming If MNO has decided to store 2-3IPDR If MNO has decided to store 0…200LA update If MNO has decided to store 12Signalling Requires additional network
hardware installation + vast storage amounts (Hadoop HDFS)
200…2000
10/06/2019 Mobile Phone Data Training 11
Data Size – 1 Year – Example Indonesia
Data type Subscribers No of records VolumeInbound 12M 1B 0.3 TBOutbound 20M 1.5B 0.5 TBDomestic 246M 1500B 500 TB
10/06/2019 Mobile Phone Data Introduction 12
ID of the Subscriber
• Four possibilities of ID:• IMSI (=>MCC)
• MSISDN (=>CC)
• IMEI (=>TAC)
• MNO customer ID
• IMSI is preferred ID as raw data
• Hashed into pseudonymous ID
10/06/2019 Mobile Phone Data Introduction 13
IMSI: 410 07 2821393853MobileCountryCode(MCC)
MobileNetworkCode(MNC)
MobileSubscription Identification Number(MSIN)
7-digit IMSI
Network Antennae Data
• Network antennae database includes:• CELL_ID• Type (2G, 3G, LTE)• Location coordinates
• Lat• Lon
• Coverage area (polygon)• Height of the cell• Power (db)• Azimuth• Angle
10/06/2019 Mobile Phone Data Introduction 15
Base Station / Base Station Controller
10/06/2019 Mobile Phone Data Introduction 16
There can be several cells in one location: E.g.: 6*2G, 2*3G, 1*LTEOne cell = one Base Station (BTS), also called node b (3G), eNB (LTE)
Base Station Controller (BSC) can control several Base Stations
Historical residual data (log files, metadata) from mobile network operators’ databases
Passive Mobile Positioning Data
Antennae Distribution: Whole Country
10/06/2019 Mobile Phone Data Introduction 18
All cells/antennae inEstonia: 36,000
Network Antennae Data
• Very often MNOs do not provide coverage areas, only coordinates (lat, lon)• Even if they provide coverage areas, these are modelled and might be
incorrect (but still better than just coordinates)• Network antennae data should be checked for incorrect values
(covered in topic 5)• If you wish for results in high level spatial resolution, then modelling
and interpolation is required (covered in topic 5)
10/06/2019 Mobile Phone Data Introduction 24
Some Descriptive Statistics – Estonia vs Indonesia
0
1
2
3
4
5
6
7
8
9
00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23
% o
f rec
ords
Hour of day
Estonia IndonesiaDifferences of record countsbetween daytime and nighttime are more noticeable in CDR data compared to signalling data.
2710/06/2019 Mobile Phone Data Introduction
0
5
10
15
20
25
30
35
40
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20+
% o
f all
subs
crib
ers
Average number of records per day
Estonia Indonesia
34.5 %
24.5 %
17 %
2810/06/2019 Mobile Phone Data Introduction
k
2k
4k
6k
8k
10k
12k
14k
16k
18k
20k
k
50k
100k
150k
200k
250k
300k
350k
400k
450k
500k
0 100 200 300 400 500 600 700 800 900
Num
ber o
f sub
scrib
ers
Time difference between records (min)
Indonesia Estonia
Location area updates for idle subscribers
The Recipe for Quality Data
Engi
neer
ing
prob
lem
-Mob
ile o
pera
tor
Data science and business problem- Data scientists and statisticians
This is where youwant to be
Activity: The metadata
http://bit.ly/mpd-jakarta
Open the mobile phone dataset provided for one person. Look through all the accompanying documents you are given. Describe the dataset in a few lines.
1. What type of mobile phone data sources do you see?
2. What is the identifier?
3. How is the timestamp structured?
4. How is the location given?
5. Which types of data are included – inbound, outbound, domestic?
6. Which types of events are included?
10/06/2019 Mobile Phone Data 30