Este documento habla sobre el uso de big data en la industria hotelera. Explica brevemente qué es big data y los diferentes tipos de datos (estructurados, no estructurados y semiestructurados). Luego describe cómo los hoteles pueden usar datos como reseñas en redes sociales, precios de la competencia, reservas, preferencias de clientes y datos meteorológicos para mejorar la toma de decisiones sobre precios y servicios. Finalmente, resume algunas de las aplicaciones prácticas de big data para los hoteles, como predecir la
Azure Data Lake: integracion dentro de soluciones de inteligencia de negocios
Presentacion big data y hoteles con tecnologia Microsoft
1. BIG DATA Y LA INDUSTRIA HOTELERA
Juan Alvarado
MVP Data Platform
2. Juan Alvarado
MVP Data Platform
.
25 años como DBA y Developer BI
Especialista en analisis Avanzado con Azure,
Big data, Power BI
/juan.m.alvarado
@juanbizzz
Juan Alvarado
juanbizzz@outlook.com
Medium.com/@juanbizzz
Consultor de SAP Hana , SAP ASE y
Oracle Essbase
3. Que es Big Data?
También llamados macrodatos o datos masivos, se le llama así a los
conjuntos de datos que son tan grandes que las aplicaciones informáticas
tradicionales no son suficientes para analizarlos.
1. Estructurados. Datos que tienen bien definidos su formato y longitud.
Como ejemplos, las bases de datos.
2. No estructurados. Carecen de un formato específico. Por ejemplo:
documentos PDF, archivos de audio y video, correos electrónicos.
3. Semiestructurados. Contienen marcadores para separar los diferentes
elementos que los componen. Algunos ejemplos son las hojas de cálculo
o los archivos HTML
4. Big Data = Mejores datos
Reviews &
Social Media
Competitor Pricing Data
Booking & Reservation Data
Web Shopping Regrets & Denials
Weather
Air Traffic
Traditional Revenue Management
Traditional Revenue Management
5. Big Data = Mejores datos
Reviews &
Social Media
Competitor Pricing Data
Booking & Reservation Data
Web Shopping Regrets & Denials
Weather
Air Traffic
Traditional Revenue Management
Traditional Revenue Management
6. Big Data = Mejores datos
Reviews &
Social Media
Competitor Pricing Data
Booking & Reservation Data
Web Shopping Regrets & Denials
Air Traffic
Traditional Revenue Management
Traditional Revenue Management
Weather
7.
8. Sentiment Analysis 101
Un vistazo el analisis sentimental
El analisis sentimental es el proceso de entender las emociones dentro
de un context de texto
9. Miremos un comentario de un hotel
I had a fantastic time on holiday at your resort. The
service was excellent and friendly. My family all really
enjoyed themselves.
The pool was closed, which kind of sucked though.
Hotel Feedback
10. Tomas las palabras positivas y negativas
Positivas
Good
Great
Fantastic
Excellent
Friendly
Awesome
Enjoyed
Negativas
Bad
Worse
Rubbish
Sucked
Awful
Terrible
Bogus
11. Hacer match dentro del comentario
I had a fantastic time on holiday at your resort. The
service was excellent and friendly. My family all really
enjoyed themselves.
The pool was closed, which kind of sucked though.
Hotel Feedback
15. Actividad de web site
Se proactive, no reactivo, con las tendencias de la demanda.
▍ Revise fechas de
busqueda, fechas
ocupacion, rates, tipo
de cuarto, rate,
fuente, y pais
▍ Entienda los
periodos altos de
demanda antes de
fijar promociones
17. Cuatro areas mayores de estadistica
17
Forecast mas exactos
Simple Error Mean Simple Percent Error
(MSPE)
Mean Absolute Deviation
(MAD)
Mean Absolute Percent Error
(MAPE)
18. ¿Cómo usan los hoteles el big data?
• Para saber en dónde abrir
• Para mejorar los servicios e instalaciones
• Para conocer el impacto de un cambio de tarifa antes de
implementarlo
• Para optimizar el manejo de inventario
• Sintetizar las reseñas de tus huéspedes
• Para mejorar el forecast de ocupacion
19. SmartHotel360 is a fictitious smart hospitality company showcasing the future of connected travel.
We built intelligent and personalized apps for guests, business travelers, and hotel managers. All powered
by the cloud, our best-in-class tools, our data platform, and Artificial Intelligence.
20. Their vision is to provide:
• Intelligent, conversational, and personalized apps and experiences to guests
• Modern workplace experiences and smart conference rooms for business travelers
• Real-time customer and business insights for hotel managers & investors
• Unified analytics and package deal recommendations for campaign managers.
21. z
z
Maintenance app
Z
Public Web Site
Azure Container
Registry
Sentiment App
ArchitectureDiagram
AAD B2C
Sentiment API
Suggestions API
Kubernetes Cluster in Azure Container Service
SmartHotel360 Customer App
Hotel API
z
Azure Databricks
(Package Deal
Recommendations)
Azure SQL DB R services
(Suggestions)
Azure SQL DB
(Bookings)
Azure SQL DB
(Hotels)
Azure Cosmos DB
(Twitter Sentiment
staging for apps)
Bookings API
Notifications API Tasks API
Config API
Twitter
Sentiment Analysis using
Cognitive Services
Azure Functions
Booking
Management
Azure DB
for MySQL
Reviews API
NFC
Profiles API Discounts API
[Moderator Part]
Esta sesión es presentada por Juan Alvarado. Juan es MVP de de Data Platform durante 11 años consecutivos. Consultor de tecnologias de SQL Server, Power BI, Project, Azure , Azure Data Lakes y Office 365. Certificado en SQL y Sharepoint. Consultor certificado sobre Sap HANA y Oracle. Especialidades: SQL SERVER, Business Intelligence, Powerpivot ,Project Server, Sharepoint, BPM, entre otras
[move to next slide]
Review Data: good to look at for a “thumbs up” or “down” approach. Can impact forecasts if heavy negative buzz in social spheres; can also be the opposite if stronger performance exists. E.g. Bed Bugs, or other serious social buzz could impact forecasts.
Specific points that you may go up or down per week wont have direct impact to property; cant link to specific customer segments.
Competitor pricing: if you are priced wrong in the market you can impact your demand in the market which then impacts your forecast
Macro trends help give market-level insight: air travel and weather can be huge indicators for city and regional markets
Web site activity gives huge indication into unconstrained market demand
You can look at pace going to your booking engine; Unconstrained demand: cite EDC example to say that you could have seen demand before people purchased to appropriately raise prices
Forward looking information – can’t just get this from Google analytics
Actual bookings are just “tip of the iceberg”: understanding pace and pick-up through lost business gives much better insight into anticipating compressed days
ties to rate code, room type, stay dates, search location (location demand)
wont hire a full time person to aggregate these data sources but having a system helps aggregate that data