December 10-12, 2015，Sponsored by China Computer Federation (CCF)，Organized by CCF Big Data Expert Committee，Institute of Computing Technology, Chinese Academy of Sciences、Co-organized by Beijing Zhongke Tianji Technology Co., Ltd. and CSDN，With "Data Security、In-depth analysis、Industrial Application” as the theme of the 2015 China Big Data Technology Conference (Big Data Technology Conference 2015，BDTC 2015) grand opening at the New Yunnan Crowne Plaza Hotel in Beijing。

In the transportation and tourism big data sub-forum on the third day of the 2015 China Big Data Technology Conference，From Beijing 188bet Online Sports Betting and Casino traffic operation monitoring and Dispatch Center、China Special Car、高德、 Ctrip、Tuniu、Six experts and professors from Tujia shared the practices of governments and enterprises in using big data in transportation and tourism。This forum is hosted by Dong Shixiao, deputy editor-in-chief of CSDN。

CSDN Deputy Editor  Dong Shixiao

Zhang Ke, Deputy Director of Beijing 188bet Online Sports Betting and Casino traffic operation monitoring and Scheduling Center: Construction and Application of Beijing’s Comprehensive Traffic Operation Data System

Beijing 188bet Online Sports Betting and Casino traffic operation monitoring and Scheduling Center (TOCC) was built at the end of 2010，Organization established in May 2011，Is an important part of Beijing’s comprehensive transportation coordination system。An operation monitoring system has been formed、Coordination and dispatching system、The overall business pattern integrating the information service system。

Deputy Director of Beijing 188bet Online Sports Betting and Casino traffic operation monitoring and Dispatch Center  Zhang Ke

Zhang Ke talks about Beijing’s monitoring data system for comprehensive traffic operation analysis。For huge basic data content and data volume，Focus on the needs of comprehensive traffic monitoring and operation analysis，Through the summary and sorting of the data layer，Combined data sources、Unify the data access method and design the data storage structure。Operational characteristics around urban road network areas and road segments，Constructed traffic index、Running speed、7 indicator items including congestion duration；Based on AFC and operational data，Constructing dynamic passenger flow indicators for rail transit、Rail transit passenger volume indicator、Rail train operation indicators；Via taxi GPS data and meter data，Building taxi capacity indicators、Operational Characteristics Indicators、Operational safety and regulatory indicators；Through in-depth mining of data items with strong cross-domain correlations，Proposed relevant indicators reflecting the characteristics of urban comprehensive transportation operations，Such as the proportion of rail transit passenger traffic、Bus transportation speed ratio、Customized multi-granularity composite feature extraction for comprehensive indicators such as intercity passenger volume and proportion。

The center also established a comprehensive traffic dynamic operation monitoring and analysis、Video resource management application、Provincial comprehensive 188bet Online Sports Betting and Casino traffic operation monitoring business platform that releases public information in one unified place。The video resource management application platform includes a digital video resource management platform、GIS-based video application platform、Video intelligent analysis platform based on image intelligent extraction technology。Unified public information release platform through information acquisition、Information editing、Information review、Information release、Information archive、Release channel docking、Statistical analysis realizes true integration、Multi-modal comprehensive travel services。On the basis of integrating existing online services in the transportation industry，Interprovincial Passenger Transport、Public bicycle、Rail transportation、Interprovincial Passenger Transport、Flight、Railway、Parking and other ways。

End of speech，Zhang Ke said that in the future, the Beijing 188bet Online Sports Betting and Casino traffic operation monitoring and Dispatch Center will open up shared data resources and simulation tool environment，Provide comprehensive traffic data viewing for traffic data professional institutions and personnel、Export、Analysis、Integrated support for simulation and visual display。

Li Si, chief architect of China Private Car: China Private Car spatio-temporal big data processing practice 

Li Si, chief architect of China Private Car, shared "China Private Car's spatio-temporal big data processing practice"。China Private Car is an Internet travel company established at the beginning of this year，The goal is to provide a high-end travel experience。China Private Car: Safety is the foundation，A two-pronged approach to management and technology，Manage drivers with strict systems and reduce bad driving behaviors through Internet of Vehicles monitoring，Ensure driving safety；Efficiency is the key，Improving driver efficiency through improved technology、Order fulfillment rate、Vehicle utilization rate；Growth is the goal，Quantitative analysis of user behavior，Achieving precise marketing。

Chief Architect of China Private Car  Li Si

Under overall structure，Collect data through Internet of Vehicles/OBD and process it in real time；At the same time, offline analysis machine learning forms a prediction model to schedule the service front-end、Pricing；Driver/passenger frontend request return to dispatch、In finance and other subsystems，Perform data analysis again。There is a business intelligence processing and analysis system behind it，Data is gathered in the database using MySQL，Finally generate business report。At the same time, many user access logs were collected on the front end，Form a user event stream，Facilitates analysis of user behavior。Ensuring users’ driving safety through driving behavior analysis and alarms；Fuel consumption for drivers、Improve efficiency through effective mileage analysis and other methods。Grid division in space，Grid internal supply and demand forecast by time，Improve scheduling efficiency；User data analysis platform passes user historical orders、Access Log、How to receive and use coupons、Data analysis such as social network sharing records to achieve user profiling and precision marketing。The specific construction method is:

1.Define user operations and system state changes caused by them as events；

2. Aggregate events from multiple heterogeneous data sources by user;

3.188bet app download Sort the events of the same user by the time they occurred，Become a user event stream；

4.Define a batch of basic functions (API) based on user event streams according to business needs；

5.Use Hive or MapReduce for data analysis。

Last speech，Li Si shared China Private Car’s experience in big data processing。For startups，There may be a shortage of rental computer room resources、Large-scale marketing activities can easily cause system overload and other problems。At this time, the company can cooperate closely with third-party companies and universities，At the same time, non-core components can be integrated (reducing the coupling between system components)、Large-scale marketing campaign、Third-party cooperation、Short-term computing-intensive tasks are deployed on the cloud (Iaas)，To achieve improved security、Improved reliability、Increased flexibility、The effect of controllable costs。

Fang Xing, senior expert on big data from AutoNavi: How big data guides travel

Fang Xing, a senior expert on big data from AutoNavi, gave a speech titled "How Big Data Guides Travel"。Produced by Gaode、All consumed is data。Through these data we can better understand travel、Geographic information and related services，Provide better support for future business。

Onsite，Fangxing displays domestic traffic conditions based on data analysis。Amap through real-time data processing，Real-time traffic conditions and events、Real-time traffic and navigation。Amap is a data company，All work is done around data，The lowest level and the core part of the traditional information system is geographical information。The technical structure is also the business structure including: front-end application online service，Include navigation、Positioning、Search、Navigation、Real-time traffic and more。Each service alone is a very huge system。The front end is mainly divided into several large blocks，The first big piece is the AMAP application APP。The second block is the open platform，Provide open data interface for third-party partners，Provide navigation services、Bus service、Positioning。The third piece is the Internet of Vehicles that we have recently focused on，Vehicle equipment，Built inside the car of the future。All data for the fourth China Apple map are provided by Amap；Hbase、Alibaba RDS、Cache and other technologies realize online data storage；Core data unified storage，Use Alibaba ODPS (cluster composed of 5000 nodes) to support multiple ways of SQL/MR/Spark/Storm，The platform’s data will be aggregated to ODPS every day，Then you can perform minute-level data reports、Data generation、Data Release。Entire map engine、Navigation engine、Real-time traffic engines work in this way；Also use TT production server real-time log collection。

Senior Expert of AutoNavi Traffic Big Data  Fang Xing

Based on these data，Amap has done a lot of data mining work: first, by combining the trajectory heat map with the existing road network，Publish new and expired road information；The second is by analyzing changes in traffic flow and events reported by users，Real-time monitoring of road closures and traffic incidents；The third is the analysis of historical vehicle speed，Building a road speed model，Used for traffic prediction。After these methods are associated，Credibility or quality、The accuracy will be greatly improved。With these methods，Users 188bet app will be more accurate when planning routes or determining travel time。

Amap can be oriented to traffic through data mining，Provide better services for users’ travel，The vision of the future is to realize intelligent transportation。

 Yu Lei, Senior Data Analysis Manager of Ctrip Basic Big Data: Ctrip Basic Big Data Architecture Practice

Yu Lei, senior data analysis manager of Ctrip Basic Big Data, shared "Ctrip Basic Big Data Architecture Practice"。For e-commerce companies in the OTA industry，Maybe its big data application is relatively simple。For example, Ctrip website，It has 15、16 business lines，But if you open JD.com or Tmall No. 1 store，You will find that it has more than 4,000 category items。

OTA company data has large differences in business lines、High complexity and other characteristics，The big data implementation scenario is often analysis reports、AB test platform、User Center、Personalized recommendation field、Precision operations, etc.。Currently, Ctrip has 20 BUs and corresponding data teams，But we are also faced with multiple data sources、Different standards、Difficulties in ETL process management、Problems such as difficulty in using cross-BU data at the company level。

Ctrip Basic Big Data Senior Data Analysis Manager  Yu Lei

Currently in Ctrip’s big data architecture，The big data architecture is divided into precise application layers、Generalized configuration layer、Algorithm engine layer、Log layer、Data source layer five。The data source layer is used to standardize buried point registration、Generate，And perform real-time environment verification (beforehand)、ETL data verification (in progress)、Offline verification alarm (post-event)；Use Trace MQ (Kafka, etc.) for offline operation in the log layer & Real-time data access，And open MQ、API Service、Hive table external interface；In the algorithm engine layer，Through offline user portrait、Similar related products、Rule Mining、Topic Mining (HBase、Redis) mode and real-time: user intent (HBase、Redis) analysis，and Hive、ES and other methods，Implementation model verification portrait review、Schema recognition in ETL、Test environment bloodline review；Generalize configuration layer to select human cube (advertisement、 SMS、APP push、EDM）、Personalized recommendation platform、Product selection platform and other general precision marketing platforms (under development)，Implement data export；The precise application layer is used for recommended fields and personalized advertising。

Currently Ctrip has implemented metadata management of data pivots、Data specification、Unified dimension table、Unified data interface、Standardized indicator、Secure management data governance process。 Metadata standard process: Currently, Ctrip has done some automated process management。When each team creates the data table，Schema for unified data collection，Compare with regularly stored data；Unified monitoring between ETL、Process and process、The relationship between delivery and delivery。When the schema of a data source or process may affect the work of other ETL or other teams，Can be monitored；By ancestry analysis，Notify downstream。

In QA session，Yu Lei gave a detailed answer to the burying techniques proposed by the guests and the dimensions to consider when recommending users。

Meng Jingci, Director 188bet online sports betting of Tuniu Big Data: Tuniu Tourism Big Data Application Practice

Meng Jingci, Director of Tuniu Big Data, gave a speech titled "Tuniu Tourism Big Data Application Practice"。His speech is divided into Tuniu business model、Travel Industry Characteristics 、Data solution three-part expansion。At present, Tuniu’s business model is divided into tourism e-commerce、Financial Technology、Three pieces of film and television media。At present, the most important thing is Tuniu Tourism。Tuniu first started with group self-service，Different from Ctrip’s model，Air tickets are also available now、Hotel、Tickets、Visa、wifi，But traditionally the most important things are group tours and self-help。

Director of Tuniu Big Data  Meng Jingci

The tourism industry is non-standardized、Low frequency、Strong timeliness + high customer unit price features。Travel is a non-standard product，This can generate non-standardized derived requirements: how to combine resources？Tuniu adopts the automatic packaging of products to similar products and resources to related resources in this regard、Methods of cross-selling。He talked about，Tuniu is not completely low-frequency，At least in the APP era，Should be between high and low frequencies。Tuniu has a destination resource，Recommendation system based on LBS，Can recommend to users the performances around the tour or the food, drink and fun to realize travel goes mobile。Already achieved a higher frequency in supplier shorts，Tuniu supplier platform N-Booking now has tens of thousands of registered users、More than 85% of SKUs are entered through NB、Millions of SKUs、Order NB confirmation rate is over 95%。

Next， Meng Jingci talked about the supplier financial control model，Adopt tourism platform + investment institutions and tourism suppliers to use platform data to jointly invest model。The timeliness and other characteristics of tourism products derive the need for risk management and price management，Produced corresponding risk management、Quantitative solutions for price management、New products and models。He mainly introduced optimal inventory and pricing strategies and revenue management strategies。Pre-sales: Determine the optimal amount of primary inventory to be purchased before the sales period，As a target for negotiations with suppliers，At the same time, formulate a price strategy corresponding to the final inventory obtained；On sale: after the sales period reaches a certain stage，If sales are significantly higher than pre-sale expectations，Calculate the optimal amount of secondary inventory to be replenished，and corresponding pricing strategy，If sales are significantly lower than expected，Adjust price strategy and take corresponding measures。In Revenue Management，Segment the market into business travelers and leisure travelers，Develop different pricing strategies，Differentiated push products，The specific implementation solutions are Demand Forecast and Inventory Allocation。At the end of the speech, he concluded that tourism products and financial needs should be closely integrated，In order to achieve the best revenue management。

Qin Yong, Director of Tujia BI: Business Practice of Tujia.com Data Analysis

Qin Yong, Director of Tujia BI shared the topic "Business Practice of Data Analysis on Tujia.com"。He briefly introduced Tujia as a global apartment and homestay booking platform，Covering 265 cities in the country，188bet Online Sports Betting and Casino 1038 overseas cities。The focus of his sharing this time is: the value of data lies in the combination of technology and business；In-depth development of data technology，A hundred flowers bloom in business practice。Qin Yong shared 4 cases of Tujia.com’s business practice in this speech，Discuss with you: "How data interacts with business at the commercial level"。

Tujia BI Director  Qin Yong

Compared with traditional OTA，Tujia’s business has a large proportion of vacation needs from families with more than 3 people；30 product types to choose from: Apartments、Duplex、Villa、Yachts, etc.；20% of user decision-making cycles (from browsing to purchase) exceed 1 day with three characteristics。Therefore, the need to solve vacation needs often does not have a clear destination or is relatively vague，Need to help users solve the "where to go" problem；Many product types，Need to help users determine “which house to live in”: user decision-making is slow，Need to help users speed up decision-making，Answer the three questions “How are these houses?”。

Through the data analysis of “Qunar” Case 1，Tujia has the following three characteristics:

1. At any time, intra-city booking orders account for a large proportion;

2. Most of the time, peripheral travel is second only to intra-city booking;

3.Going between cities for vacation is regular: entering winter，The largest increase in the number of people going to Sanya from the Northeast and Southwest。

Therefore, Tujia has developed three corresponding business applications:

1.According to the user’s location on the Tujia PC homepage，Made destination intelligent recommendation；

2.The smart recommendation clearly specifies the house recommendations in this city、Recommended around the city，Other city recommendations；

3. Strengthened the intra-city brand promotion of offline stores.

Case 2，Discover patterns by collecting user history data。Based on the summarized rules，Give each house a suitable travel purpose，Easy for users to choose。Case 3，Enhancing user information requires some emotional evaluation: read the comments，But there are two problems: there are too many comments，And usually in the later process (details page)。By text mining，summarizes the impression of each house，And arrived at the list page display in advance。

These cases closely focus on Tujia business，Design calculation method，Output analysis results、And implement it to drive business improvements，Qin Yong gave three inspirations: First, the starting point of data analysis must be related to the specific business，Such as measurement business and prediction business；

Second, there is no fixed method for data analysis，Measurement business is often a simple calculation based on a deep understanding of the business，Mathematical models are often used in business forecasting；Third, the rotation of data analysts in business units improves the value of data analysis results、An effective way to strengthen the interaction between data and business。

188bet Online Sports Betting and Casino

In addition to sharing useful information，Zhihui Travel will also draw lucky visitors on site to give away Seychelles provided by Anke Yunda、Bali round-trip air ticket！Hurry up and make an appointment！Please click ☞ to register“I want to register”！