Building a tourism intelligence system using big data

Building a tourism intelligence system using big data Jon Kepa Gerrikagoitia, Ph. D. OPTIMA / Optimization Modelling & Analytics ICT - European Software Institute Division jonkepa. gerrikagoitia@tecnalia. com Aurkene Alzua Sorzabal, Ph. D. CICtour. GUNE Aurkene. Alzua@tourgune. org

The importance of tourism worldwide Tourism is an economic phenomenon concerning the movement of people to places outside their usual environment for either personal or professional purposes. Great impact of tourism on the economy (employment, income): need for detailed and updated information in “real time” employment GDP UNWTO (2014)

Our challenge Incorporate new measurement methods based on digital footprint (Big Data) as an alternative and complementary source in the statistical production Why: Efficiency (cost, time) How : Tourism Observatory as intelligence platform in tourism to leverage innovative business models based on data economy 3

Tourism Observatories – conceptual framework The Relationship between statistics and technology 4

Tourism Observatories –sources and methods

Tourism Observatories – technology

Tourism Observatories – types of analytics

Dynamic Pricing monitor - Big data and its uses in the tourism statistics • Official figures are published with a delay up to weeks. The system can publish the data at the end of the collection period (for example the last day of the month) • Good fit and strong correlation between pricing official statistics and the monitor’s data. Difference aprox. 5% • Modelling and prediction of hotel occupancy rates based on room prices offered online • Dynamic pricing patterns and impact 8

Dynamic Pricing monitor - Big data and its uses in the tourism statistics Hotel occupancy estimates at subnational level, Basque Country (Spain) • Data: 1. 984. 149 observations for the Basque Country from 2013 -01 -01 to 2013 -12 -31 • Method: The system collects double room price 1, . . . , 28, 45, 60, and 90 days in advance of the target date. For each hotel, three percentiles (P , and P ) have been used as the bounds: 30 70 99 • Seg 1 low prices • Seg 2 normal or middle prices • Seg 3 high prices • Seg 4 unusually high prices • Results: Fitted linear regression model Model stimates and residuals 9

Dynamic Pricing monitor - Big data and its uses in the tourism statistics Occupancy prediction model at subnational level, Basque Country (Spain) • Model that allows to predict occupancy rates before the official figures are available • Segmentation bounds obtained for the 2013 data (training set) and model are applied to data for 2014 (test set) Predictions for the Basque Country for 2014 10

Dynamic Pricing monitor - Big data and its uses in the tourism statistics Price, seasonality and trend Spain, Basque Contry, San Sebastian and Rioja in 2014 11

Dynamic Pricing monitor - Big data and its uses in the tourism statistics Destination performance analysis - Bilbao 12

Dynamic Pricing monitor - Big data and its uses in the tourism statistics Benchmarking Regions in Spain 13

Dynamic Pricing monitor - Big data and its uses in the tourism statistics Benchmarking Regions in Europe (France, Spain, Ireland) 14

Dynamic Pricing monitor – Benchmarking Countries (France, Spain, Ireland) 15

Dynamic Pricing monitor Benchmarking cities (Bilbao, San Sebastian, Vitoria) June – September 2013 June – September 2014 16

Dynamic Pricing monitor - Big data and its uses in the tourism statistics Informal accommodation monitoring - Airbnb - July 2015 Apartment - w (N=622) Mean price (€) room Araba Bizkaia Gipuzkoa 35 54, 1 60, 2 119 173, 5 apartment 130 airbnb offer in july in the Basque Country Araba 2% Private room - p (N=437) Bizkaia 32% 17 Gipuzkoa 66%

Social Media Monitor • • Monitoring platform and active listening in social media that allows Learn and discover: what (concepts) - how (polarity) - where (sources of opinion) and management: early warning, real-time monitoring online reputation using language processing technologies, visualization and analytical modeling It has been applied in the fields of Tourism, Culture, Territory, Transport and Innovation models.

Destination Image Framework

Results

Results

Destination Web Monitor (DWM) is “a system to measure, analyse, and model the behaviour of visitors in different virtual areas in which a destination is promoted and with the objective of providing benchmarking ratios that facilitate strategic surveillance and intelligent marketing policies”

Destination Web Monitor Descriptive analysis Typologies Navigation patterns

Thank you ! Eskerrik asko! Muchas gracias ! Jon Kepa Gerrikagoitia, Ph. D. OPTIMA / Optimization Modelling & Analytics ICT - European Software Institute Division jonkepa. gerrikagoitia@tecnalia. com Aurkene Alzua Sorzabal, Ph. D. CICtour. GUNE Aurkene. Alzua@tourgune. org