Shirin Najafabadi
The City University of New York (CUNY)
Projects
Quantifying the Impacts of Rainfall on Subway Ridership in Manhattan
The Northeast United States, particularly New York State has experienced an increase in extreme daily precipitation during the past 50 years. Recent events such as Hurricane Irene and Superstorm Sandy, have revealed vulnerability to the intense precipitation within the transportation sector. In the scale of New York City, where transit system is the most dominant mode of transportation and daily mobility of millions of passengers depends on it, any disruption in the transit service would result in gridlocks and massive delays. To assess the impacts of rainfall on the subway ridership, we merged high-resolution radar rainfall and subway ridership data to conduct a detailed analysis for each of the 116 subway stations at the borough of Manhattan. The analysis is carried out on both hourly and daily resolution level, where a spatial-temporal Bayesian multi-level regression model is used to capture the underlying dependency between the parameters. The estimation results are obtained through the Markov Chain Monte Carlo sampling method. The results for daily analysis indicate that during weekdays, transit ridership in the stations located in commercial zones is less sensitive to the rainfall compared to the one with residential land use.
Methods: Hierarchical Bayesian Regression Model, Markov Chain Monte Carlo Sampling Method, Statistical Hypothesis Analysis, and Bootstrapping.
Download Poster
Integrated Approach for City Logistics: A Dynamic Taxi Crowdshipping Model
The rapid evolution of e-commerce and online shopping has led to many challenges for the movement of goods within the city. Advanced integrated city logistics approaches are desired to ensure efficiency and reliability of urban mobility for both people and goods. This paper contributes to the use of existing urban infrastructure for serving parcel and passenger simultaneously. We proposed a taxi crowdshipping model that leverages the parcel deliveries by occupied taxis while serving passengers. This paper considers the problem of matching taxis and delivery requests in a real-time dynamic setting. A Mixed Integer Linear Programming (MILP) formulation is presented and used within a rolling horizon scheme that periodically updates input data information and designed to handle uncertainty caused by the dynamic nature of this problem. The aim of the study is to minimize the total system-wide vehicle miles incurred by system users, individual travel costs and a number of delivery vehicles. The results show that the proposed dynamic model is promising and has great social advantages such as saving travel cost, reducing travel time and mitigating traffic congestion.
Methods: Mixed Integer Linear Programming (MILP), Gaussian Copula Method, and Rolling Horizon Strategy
Crowdshipping Survey
Crowdshopping is innovative and can be a potential solution for last mile delivery problem. In crowdshipping,
ordinary people, rather than professional service providers, deliver goods. We conducted a survey as part of this academic research project to better understand peoples' attitude toward crowdshipping idea. This survey should take less than 10 minutes to complete. I really appreciate if you can participate in this survey.
Survey Link: https://www.surveymonkey.com/r/ccnycrowdshipping2017
Inference of Pattern Variation of Taxi Ridership Using Deep Learning Methods: A Case study of New York City
Taxis constitute an important component of the public transportation infrastructure in large metropolitan areas. However, when seen within a supply and demand framework the operation of taxi transportation system is far away from its optimal equilibrium, yielding a missed cost of opportunity for customers, drivers and city planners. The key for optimizing its market lies in forecasting taxi demand with high geospatial-temporal precision. In this paper, the taxi pickup pattern is predicted by utilizing a deep learning approach that leverages Long Short-Term Memory (LSTM) neural networks. This study is based on publicly available taxi data for the New York City. Pickup data are binned based on geospatial and temporal informational tags, which are then clustered using Principal Component Analysis. The spatiotemporal distribution of the taxi pickup demand is studied within short-term periods (next one hour) as well as long-term periods (next 48 hours) within each cluster. The performance and robustness of the proposed technique are evaluated through a comparison with Adaptive Boosting and Decision Tree Regression models fitted to the same dataset. Numerical results show the dominance of the LSTM model on the short-term horizon and relatively smaller errors for the long-term prediction.
Methods: Principal Component Analysis, AdaBoost Regression Model, Decision Tree Regression Model, Deep Learning- Long Short-Term Memory (LSTM) Model, and Neural Networks.
Download Poster
New York City Taxi trip Records: Trend Analysis and Pattern Recognition
Today, traditional ways of freight delivery and in general city logistics in dense urban areas are challenging. Therefore, providing enhanced shipping system requires a better understanding of trip patterns, integration of different modes to improve accessibility and better management of the transportation system to accommodate demands. Different factors affect city logistics such as geographic, economic, social, and cultural circumstances and attitudes towards city logistics vary in the different parts of the word. There are many trends affecting city logistics such as urban population growth, e-commerce development, the desire to expedite service and sharing economy.
The purpose of this study is to investigate taxi trip data and evaluate a model to predict taxi trips. The study area is the five boroughs of New York City. To accomplish this goal, first I used grid decomposition method to classify the study area into equal size grids and each grid considered as a node. Then the spatial and temporal variation of taxi trips in New York City (NYC) is explored using different machine learning and pattern recognition methods such as Principal Component Analysis (PCA), Trend Analysis and Wavelet Transform method. Then using the results from previous parts and based on the real taxi trip data in New York City the model is developed. The preliminary results of the model are very promising and indicate a high r-squared value of about 89 %.
​
Methods: Grid Decomposition, Descriptive Data Analysis, Augmented Dickey-Fuller Test, Principle Component Analysis, Seasonality Decomposition Analysis, and Wavelet Transform.
​
Vertical Integration of Regional Transportation Plans and Land Use using Node-Place Approach; Case Study of I-287 Corridor
This is a report completed as part of a research project and published on the University Transportation Research Center (UTRC - CCNY). The project description and scopes are listed:
Project Description
New York Metropolitan Transportation Council's (NYMTC) members introduced the idea of Coordinated Development Emphasis Areas (CDEAs) in Plan 2045, the recently adopted Regional Transportation Plan. These are areas within the NYMTC planning area, where land development and transportation investment planning are to be coordinated to achieve environmental sustainability, local economic revitalization, and improved quality of life. This coordination will challenge NYMTC as land use decisions are controlled by the 202 local municipalities within its planning area. The researcher will research methods to enhance the influence of NYMTC’s regional transportation plans on municipal land use planning decisions. Also the researcher will investigate methods to help NYMTC ensure that municipal planning efforts are incorporated into the regional planning perspective.
Project Scopes
1- Literature Review & State of the Knowledge
The researcher will review relevant information as a basis for the project. Documentation to be reviewed will include NYMTC’s Compendium of Agreements and Operating Procedures, the land use-transportation sections of the current NYMTC plan and its three most recent predecessors, the regional sustainability planning section of the New York-Connecticut Sustainable Communities Initiative, research on the topic performed by the National Center for Sustainable Transportation, National Association of Regional Councils, Association of Metropolitan Planning Organizations, American Planning Association and the Transportation Research Board, and the relevant sections of New York State law governing land use powers and federal regulations governing metropolitan transportation planning.
2- Benchmarking
The researcher will research the land use-transportation planning practices of peer metropolitan planning organizations (MPOs) in northern New Jersey, Philadelphia, Chicago, Detroit, Denver, San Diego, Los Angeles, San Francisco and Portland, Oregon and develop a comparative analysis of the practices of these MPOs.
3- Information Gathering
The researcher will collect the relevant master plans and/or zoning ordinances from the 202 local municipalities and five suburban counties in the NYMTC planning area and catalog access to these documents and the contact information of the relevant municipalities/counties.
4- Synthesis
The researcher will analyze up to ten CDEAs in NYMTC’s Plan 2045 relative to the relevant local master planning and zoning. The researcher will then apply the research from the literature review and benchmarking to develop recommendations for each of these test areas. When completed, the researcher will draw inferences from this synthesis and develop programmatic recommendations on the topic for planning practices that could be employed by NYMTC to foster the vertical integration of transportation and land use planning.
Click here to read about the presentation:
Click here to download the presentation:
Vertical Integration of Land Use and Transportation Planning
Click here to view the meeting:
NYMTC’s YouTube page
Additional photos are available on
NYMTC’s Flickr Page.
