ENTER2018 eTourism Conference presentation

1. Big data in crisis
monitoring:
case study Flanders and
Brussels
Steven Valcke
Visit Flanders
IFITT 2018, Jönköping

2. Flanders
Northern part of Belgium
Visit Flanders is responsible for promotion of and
destination development in Flanders and Brussels
A research team of 7
One of 3 Belgian NTO’s

3. Big data at Visit Flanders
• 5 years+ of experience and experimenting with
ups and downs
• Only ready for use by third parties (we are not
the data scientists)
• Sources:
1.Credit card data for expenditures
2.Mobile data for measuring the success for
certain events, on ad hoc basis;
3.webmonitoring (all public available sources
incl social media) for social listening,
brand monitoring and campaign monitoring;
4.Flight data to monitor campaigns and do
research on booking behaviour;
5.Scraping hotel review scores to monitor the
satisfaction of the visitors

4. Start
Control
reputation
Detect
influencers
Monitor
competitors
Find customer
insights
Engage on major
social networks
Prevent Bad Buzz
Listening is owned by PR team
worried about the harm of a
potential social media crisis
Measure Campaigns
Marketing teams use social
media to fuel campaigns and
measure results with various
metrics (#fans, likes, sentiment,
etc.)
So What?
Brands struggle to leverage
the value of listening and
engagement
Social Becomes Strategic
Top management set
objectives and clear
performance indicators to track
marketing and PR results
Build
Benchmark
s
Incentivize
teams
Implement online
customer support
Build the Voice
of the Customer
Connect social
media results
with customer
surveys to get
the full picture
of the customer
journey
Integrate with Sales
Compare social media
indicators with customer
satisfaction metrics (Net
Promoter Score) or sales
Scale across
the enterprise
Predictive analytics
THE SOCIAL INTELLIGENCE MATURITY CURVE
MAKE SMARTER BUSINESS DECISION WITH SOCIAL INSIGHTS.

5. Attacks Brussels 22nd March 2016

6. Implications tourism industry
• Tourism industry was hit hard
• Damage: brand and numbers of tourists
 Crisis management
 Market specific communication required
 Big data leaded the way we communicated and
which actions were taken
 Monitoring:
 Brand damage: web monitoring by Synthesio
 Numbers of tourists: flight data by Amadeus

7. Webmonitoring – Query setup
• Query from the Paris attacks Nov 15 was
reused
• Query Requirements:
• Built fast
• Languages of all of our markets
• Tracking of unsafety feeling against
our destination
• Query NON requirements
• Track all news
 Simple query
 Lot of noise
("flanders" OR "flemish" OR
"belgium" OR "belgian" OR
"mechelen" OR "antwerp" OR
"bruges" OR "brussels" OR
"leuven" OR "ghent")
AND ("unsafe" OR "safety" OR
"cautious" OR "anxious" OR
"dangerous)
in English, French, German,
Dutch, Spanish, Italian,
Swedish, Norwegian, Danish,

8. Webmonitoring - interpretation
• Tracking of volume (and sentiment):
• Challenge: Determining the level of noise:
Average of volume (excl outliers) of the week
before
+ or –
Standard deviation
• Normal situation: three days in a row back to
noise level

9. Webmonitoring - results
-
10,000
20,000
30,000
40,000
50,000
60,000
70,000
80,000
day -
1
day 0day 1dag 2day 3day 4day 5day 6day 7day 8day 9 day
10
day
11
day
12
day
13
day
14
day
15
day
16
day
17
Numberofmentions
Total buzz
March 16 November 15
Source: Synthesio
-
20,000
40,000
60,000
80,000
day 0 day 1 day 2 day 3 day 4 day 5 day 6 day 7 day 8 day 9
Other Professional news, Industry news
Forum National newspaper
Blog TV, radio
Regional newspaper General news, Magazine

10. Webmonitoring - results
0
500
1000
1500
2000
2500
3000
Numberofmentions
Spain
0
50
100
150
200
250
Numberofmentions
Russia
Source: Synthesio

11. Flight data
• We used:
• Data on weekly base
• Only leisure
• Natural change of number of bookings throughout
the year!
• Only flights
 No indications for neighbouring countries

12. Flight data - interpretation
Combination of two approaches:
1. Year over year (15 vs 16):
Ratio n week xy / n week xy-1
Eliminate natural booking pattern
2. Within year (fix week 3):
Ratio n week x / n week 3
Eliminate change year over year

13. Flight data - results
0.00
0.20
0.40
0.60
0.80
1.00
1.20
1.40
13 14 15 16 17 18 19 20 21 22 23 24 25
Week number
Spain
fix week 3
15 vs 16
0.00
0.20
0.40
0.60
0.80
1.00
1.20
1.40
1.60
13 14 15 16 17 18 19 20 21 22 23 24 25
Week number
Russia
fix week 3
15 vs 16
Source: Amadeus

14. Learnings
1. Big data can lead the operations
2. Experience with the data is required
3. Create references to make correct conclusions
4. Queries don’t need to be complex
5. Markets react very differently in case of
crisis situations