INDONESIA’S EXPERIENCE OF USING SIGNALING MOBILE POSITIONING DATA FOR OFFICIAL TOURISM STATISTICS STATISTICS INDONESIA By Titi Kanti Lestari, Siim Esko, Erki Saluveer, Sarpono and Rifa Rufiadi (BPS – Statistics Indonesia and Positium LBS)
INDONESIA’S EXPERIENCE OF USING SIGNALING MOBILE POSITIONING DATA
FOR OFFICIAL TOURISM STATISTICS
STATISTICS INDONESIA
By
Titi Kanti Lestari, Siim Esko, Erki Saluveer, Sarpono and Rifa Rufiadi
(BPS – Statistics Indonesia and Positium LBS)
STATISTICS INDONESIA
2
INDONESIA AND EUROPE
STATISTICS INDONESIA
3
INDONESIA AND NORTH AMERICA
STATISTICS INDONESIA
BPS rely on the Immigration Record and Border Survey
for Inbound & Outbound Tourism Data
Visits from neighbouring countries only 7% of tourism
Under Coverage :
Not All Border Gates have 24/7 Immigration service
Not All Borders have border gates
Border Survey is too Expensive
4
BACKGROUND
5
Riau islands (SG border) Borneo (MY)
Sintang&Talaud (PH)
Papua (PNG)
East Nusa Tenggara (TL)
Malaysia
Indonesia
Immigration entry points
100 km
Ka
lima
nta
n -
Born
eo
Roamer trip start areahigh low
STATISTICS INDONESIA
7
ALIGNMENT WITH STRATEGY
Big Data- Can be compiled automatically- Real Time- Less Manual Labour
The integration of these new data with traditional data to produce high-quality information that is more detailed, timely, and relevant for many purposes and users, especially to foster and monitor sustainable development.
To measure the mobility of people, including mobility of tourists. The digital footprint left by the users is very sensitive, but also highly valuable, as it provides new possibilities to measure and monitor the spatio-temporal activities of the population.
In line with UN Recommendation “A World that Count”
Data Revolution For Sustainable Development
MPD as one of the Most Promising ICT Data Sources
STATISTICS INDONESIA
CDR – records of active use of mobile phones
Signalling – all signals of mobile phones captured
by the BTS (mobile antenna), even with no call/text
Indonesia uses signaling data due to its precision -
it captures more trips than CDR.
• Signaling data detected more roamers than CDR
(on average 3.47 times).
• So, it overcomes undercoverage issues,
especially in less mobile BTS-dense area.
8
SIGNALLING DATA CAPTURES MORE
STATISTICS INDONESIA
Natuna island
Malaka on the Timor Leste border
Anambas islands
Kupang on the Timor Leste border
Talaud islands
Sangihe islands
Boven Digoel near Papua
1
10
100
1000
1 10 100 1 000 10 000 100 000
Nu
mb
er
of
trip
s fr
om
sig
nal
ling
dat
a (m
ult
iplie
r ra
tio
)
Number of trips from CDR data
Signalling contributes most in hard-to-reach areas
9
SIGNALLING DATA VS CDR
More trips
Even
more
tri
ps
from
sig
naling
da
ta
1. Islands
2. Border to less
developed countries
STATISTICS INDONESIA
10
ROAMING≠TOURISM≠IMMIGRATION
0
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
MPD vs IMMIGRATION, BINTAN, JULY 2017
Immigration MPD
1 456
4384
-
2 000
4 000
6 000
8 000
10 000
12 000
14 000
16 000
18 000
20 000
Total
BINTAN, JULY 1
IMMIGR MPD
618
17624
-
2 000
4 000
6 000
8 000
10 000
12 000
14 000
16 000
18 000
20 000
Total
BINTAN, JULY 23
IMMIGR MPD
?
STATISTICS INDONESIA
Statistical and Non Statistical Noise
Fast fliers
Seamen
Accidental roamers
Other transit
11
SIGNALING DATA ADDS NOISE
MPD
IMMIGRATION
STATISTICS INDONESIA
13
MPD ERROR
0%
600
%1
200% Total improvement in RMSE
0
1
2
3
4
5
6
Error: MPD original Error: LOS1 Error : LOS1 + Seamen excluded Error: LOS1 + Seamen+Fastflyexcluded
RM
SE in
th
ou
san
ds
MPD Error (RMSE) decreases significantly after adjustments
11
x
5x
imp
rove
me
nt
8x
MPD all trips in Bintan vs main port immigration data, rough adjustment with Positium methods
STATISTICS INDONESIA
14
IMMIGRATION DATA VS MPD
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
With and without filters: Tourism arrivals in Bintan
MPD original MPD wo LOS1+seamen+fast flyers Immigration
STATISTICS INDONESIA
Implement MPD Data since October 2016.
Data provided by MNO aggregate table.
Oct-Dec 2016 : MPD was applied to 19 districts,
since January 2017 became 25 districts.
Filtering using a simple 7/20 formula (until Dec 2017), neighbouring countries.
Since January 2018, Filtering MPD has already been using a new formula
Now BPS has an MoU with the biggest MNO in Indonesia, so we can have a
sandbox to process MPD
From 2019, automatic system of processing tourism statistics with all the
necessary algorithms to remove noise
15
CONTINUOUS IMPROVEMENT
STATISTICS INDONESIA
16
MAKING OF
STATISTICS INDONESIA
𝑨𝑻 = 𝑴𝑷𝑫
𝑿𝒓𝒐𝒂𝒎∗ 𝑷 ∗ 𝟏 − 𝑷𝒘 ∗
𝟏
𝟏 − 𝑷𝒏𝒓 − 𝑾𝑪𝑰
FRAMEWORK QUESTIONNAIRE
EXTRAPOLATION FORMULA
STATISTICS INDONESIA
18
EUREKA MOMENT
STATISTICS INDONESIA
19
RESULTS
Number of additional Tourists Jan-July 2018
MPD significantly increases the coverage of inbound tourism in Indonesia.
STATISTICS INDONESIA
20
Indonesia Other countries
7% 30-60%
Before MPD use (2015)
Indonesia Other countries
7% 30% 30-60%
After MPD use (2018)
STATISTICS INDONESIA
21
HANDBOOKS
Methodology Concept and QAF Handbook
Indonesia currently already has draft of Methodology Concept and QAF handbook for the use MPD in Cross Border Inbound Tourism
Standard Framework
GSBPM
Data Collection
Data Processing
System Design
Data Analysis
Methodology Concept
Standard Framework
GSBPM
Quality Gates
Input Quality
QAF Dimension
Output Quality
QAF
THANK YOU
STATISTICS INDONESIA