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

https://unstats.un.org/bigdata/taskteams/mobilephone/Handbook%20on%20Mobile%20Phone%20Data%20for%20official%20statistics%20-%20Draft%20Nov%202017.pdf

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

Mobile Positioning Data

410/06/2019

Forms of Data


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.


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


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


Example of CDR data8 location events over 3 days

Example of Signalling Data289 location events over 3 days(but many events in the same location)

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


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


IMSI: 410 07 2821393853MobileCountryCode(MCC)

MobileNetworkCode(MNC)

MobileSubscription Identification Number(MSIN)

7-digit IMSI

Mobile network antenna

10/06/2019 14

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


Base Station / Base Station Controller


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


All cells/antennae inEstonia: 36,000

Antennae Distribution: Rural Areas


37km

Antennae Distribution: Urban Areas


Antennae Distribution: Urban Areas


200m

Antennae Coverage: Urban Areas


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)



Estonia = CDRIndonesia = SignallingGeorgia = CDR

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.


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 %


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

http://bit.ly/mpd-jakarta

What is mobile phone data? - ESCAP

Documents

Transcript of What is mobile phone data? - ESCAP