We fused data science and advanced application to create this “4D Microscope of Taipei Public Transit”. Set the location by dragging the GIS pin to any place on the map. With the multi-options searching design, including the time slot (Ex: weekend & weekday), departure time, walking ability and time-based search, this microscope will show you the comprehensive topology map of service scope in all kinds of transportation systems and Time&Space variations.
Data Mining and Forecasting of Traffic Accident Data
Applying data mining methods to analysis traffic accident data in Taipei City, then proposing accident improving strategies to reduce the number of traffic accidents base on the analyzing results, also developing accident GIS website of Taipei City to display accident thermodynamic diagram under various accident attribute condition, the important contents of project are as followed:
In order to reduce the influence generated by non-correct recorded traffic accident data, pre-processing methods of traffic accident data is proposed, including checking duplicated records, interpolating missing data, checking the currency of range and processing coordinates which are deviated from the road.
Data mining including two section which are attribute exploration and space exploration, attribute exploration analyzing the significant environment variables of traffic accidents by using the cluster algorithm.
Space exploration is done by dividing Taipei City into several grids and forecasting the amount of A1, A2, A3 traffic accidents respectively, applying artificial neural network(ANN) to predict the potential location of traffic accident and analyzing the accident properties in every grid.
The proposed data mining model can be applied to analyze notable characteristics of traffic accident, explaining the relationship between traffic accident and environment variables, then adopting accident prevention actions for specific groups.
Data Analysis and Visualization System for Traffic Management
The system can help the related sectors in the city government to discover traffic issues and making decisions.
To enhance the quality of traffic, we’ve helped Taipei City Government create a data analysis and visualization system based on various kinds of historical data. The system has three topics: Traffic Flow, Traffic Accident and Parking, with the most important applications as follows:
To predict roads or areas tend to congestion in certain period of time or circumstances.
To observe VD(Vehicle Detector) data with visualized interface to discover a certain period of time when traffic tends to congestion or abnormality.
To discover traffic accident hotspots。
To find out the relevance between accident repeaters and violations for the related sectors to take preventive actions in advance.
To observe the data of parking lots with visualized interface to find out the attributes (type of users, parking time, etc.) in different areas or in certain period of time.
Traffic Decision Support System for Taoyuan Government
We analyzed three topics of transportation data:
To forecast the velocity of vehicles on the road by appling the KNN algorithm to the data from Vehicle Detector.
To analyze and visualize the mobility and service level of public transit under different time&space condition
With the data from the eTag system, we analyzed the travel time between different road section, and found out the potential site to install the eTag deive.
Public Transport data eXchange-Design of Data Model
Those who own big data lead the growing momentum of the next generation. To own big data successfully, it relies on the design of data model, relevant to quality and quantity of data.As a team dedicating to promoting the engine of public transport data, and with the leadership of Dr. Chueh, we worked with our partners in MOTC (Ministry of Transportation And Communications R.O.C.) to the develop data model for public transport in Taiwan. Through our effort, the data model incorporates data from bus, railway, bike, and aviation, and meets the consistency and extension of data as well as requirements of international data exchange standard. Based on this, we provide a sharing platform for all VASP (Value-Added Service Provider) and startups to accelerate their innovation in intelligent transport and smart city
Complete Data Incorporation: With more than 100 data types and 600 data field across bus
Consideration of Local Operation: Demonstrate the diversified operations of railway and bus
Provision of Dynamic Data: Access to real-time bus data