I am Mustafa Fardin,
a Public and Freight Transportation Enthusiast from Bangladesh.
I am Mustafa Fardin,
a Public and Freight Transportation Enthusiast from Bangladesh.
Embedded Files
a Public and Freight Transportation Enthusiast from Bangladesh.
A first year PhD student at the Pennsylvania State University with research interests in Public and Freight Transportation Systems & Operations, while employing Machine Learning techniques along the way. Choice models and behavioral research are another key area of interest for me, while my enthusiasm for a better world poise me to integrate sustainability and equitability into my works and interests. From experiences working in various research and personal projects while also working as a Lecturer at Dept. of Civil Engineering of Prime University, I have developed my knowledge and skill in various aspects of transportation engineering; however, I particularly love working with data. Scroll below to find out more!
Research
Research
Identifying the Change in Mode Choice Behavior Due to COVID-19 Using Multinomial Logit Models & Interaction Analysis
My undergrad thesis examined the changes incurred by COVID-19 to mode choice behavior, trip frequency and change in attitude towards trip and vehicle characteristics before making a mode choice decision. Shift in modal shares and attitudes were tested using non-parametric tests such as, Wilcoxon's signed rank test, McNemar's test etc. After confirmation of the shift, the behaviors were modeled using multinomial logit models corresponding to before and after covid-19. The shift in parameters were further confirmed using interaction effects of time period.
Expand for more details regarding the thesis ⇣
Expand for more details regarding the thesis ⇣
Figure: Methodology of The Study
Abstract
Various measures, recommended or imposed by the Government of Bangladesh to control the spread of COVID-19, impacted the travel behavior of the urban travelers significantly. Showcasing high impacts while the lockdowns in place, similar research suggested that the effects of such will impact mode choice behavior significantly in a period after lockdowns as people blend in with new normalcy. This study aimed at examining the changes that occurred in travel and mode choice behavior after the repealing of lockdowns and models the change in the factors that affect mode choice. Data were collected through an online questionnaire survey, limited to Dhaka Metropolitan Area and dispersed through social media ad campaign, which included questions on trip purpose, mode choice, distance traveled, and frequency of trips before and after COVID-19 lockdowns. Two hundred and ninety-five responses were collected from the survey which were analyzed through non-parametric and parametric analysis to compare between pre-lockdown and post-lockdown periods.
Processing and checking suitability of the data for models, Multinomial Logit Model was deemed representative of the behavior and different models for Before and After Lockdown situation, considering 3 different aggregation of modes were estimated (Public Transport was used as reference). Qualitative interpretation of various factors and their impacts on mode choice behavior was undertaken. As MNL does not allow for comparing between two different time periods, Interaction analysis was done, with the introduction of a time period dummy variable. Interaction of the dummy variable with other features were calculated and interaction model was estimated. It provided insights on which changes were truly significant, leading to steps and policy reformations to be implemented in case of a future lockdown/pandemic situation, so that public transport does not decline.
Results explained that trip purpose, mode choice, cost of trip, frequency of trips and importance imparted on various Levels of Service (LOS) factors for the primary travel were significantly different before and after the lockdown. There was a significant shift from public transport to private transport and non-motorized modes while people placed a higher priority on the pandemic related concerns while choosing a mode during the pandemic as compared to the general concerns or trip attributes. Importance imparted on various vehicle characteristics, trip attributes and safety related attributes shifted significantly as travelers preferred shorter distance, shorter time and cost-efficient modes in pre-lockdown time, while preferring safer, higher cost, less distance and crowding to be more important factors in the new normal situation. Significant changes in factors affecting mode choice were observed, as household income, employment status and safety & distancing in vehicle impacted travelers to shift from public transport to private vehicles and non-motorized vehicles. Government authorities could utilize such knowledge for planning smart and partial lockdowns in the event of future lockdowns and also, utilize the knowledge in making the new public transportation systems to match users’ needs.
Sample Results
Outcomes
Successfully identified and interpreted implication of safety and health related factors of vehicles on urban traveler’s behaviors and what makes them shift mode choice in new normalcy. Also, quantifiably identified shift in modal split in new normalcy and which key points are important to focus on to regain modal share for public transport.
Discerned focal points to prioritize on in case of a new pandemic/lockdowns and recommended policy reformation to not lose modal share of public transports.
With Dhaka being the capital of a developing country, with various transportation projects being undertaken right now, forecasting use and feasibility analysis of future projects need to be reconsidered. Thus, the study opened a new door to for health related & psychological factors to be considered for the planning of future modes and such.
Traffic Volume and Peak Flow Study of Russel Square to Panthapath Corridor
The purpose of this study determined the traffic volume & flow characteristics of the road near Panthapath intersection. The objective of this study is multifaceted, in the sense of having specific objectives and general objectives which lead to the scope of various design, improvement related findings. A preliminary survey was conducted to choose the type of vehicles & their approximate ratio. Then the vehicle count was found by manual methods and manual equipment were used. The result found in this survey was documented and percentages were found for other study procedures. The data collected from all the groups were used to find composition, direction of flow, PHF (Peak Hour Factor).
Traffic Speed-Flow Characteristics of a Busy Dhaka Metropolitan Area Corridor
The proposed look into work goes from breaking down the modular speed of vehicles, highway outline component, upper and lower speed restrain for regulation, planning and analysis, traffic operation, control and direction. The range, time and conditions of the study may be coordinated by its objective and expansion. In this study, we conducted a speed study on the road between Panthapath Signal and Russell square. Manual method was used due to lack of resources and simplicity of operation. Various spot speed measures were obtained, which showed that free flow condition was not present on the corridor. The Wardrop relationship was established and evaluated.
Expand for more details regarding the two studies ⇣
Figure: Methodology of The Study
Abstract
For the completion of CE 454 - Transportation Engineering Sessional II Term Paper, we were tasked with analyzing the characteristics of the Russel Square - Panthapath Intersection corridor for traffic volume, speed-flow and potential bottlenecks were analyzed using simulation in PTV VISSIM. Statistical Analysis was conducted and the characteristics were quantified and compared with standards.
For Traffic Volume Analysis The data collected from all the groups were used to find composition, Design Flow Rate direction of flow, PHF (Peak Hour Factor). With the help of the data collection, an attempt had been made to understand the traffic patterns during different time periods. For Speed-Flow Characteristics Analysis, Maximum vehicle speed was found to vary between 15-20 kmph in Russell Square to Panthapath direction and 30-35 kmph in Panthapath to Russell Square. 50% of the observed vehicles were light vehicles. Various spot speed measures were obtained, which showed that free flow condition was not present on the corridor.
The frequency distribution curves in each direction were nearly symmetric along axis. Space mean speed was found to be less than Time mean speed. The Wardrop relationship was established and evaluated. Level of Service was found to be LOS F. Yearly cost in delay was found to be 33385971 BDT, which is a loss incurred on the country’s budget. Thus, a lane expansion project was evaluated to be beneficial for a 20-year design period. Furthermore, various other protective and control measures were recommended for betterment of the speed characteristics of the vehicles on the road.
Sample Results
Outcomes
Determined the current state of speed-flow characteristics of the corridor and compared against baseline for Dhaka City. It broadened perspective on traffic analysis and methods utilized for such.
Identification of bottlenecks on the road and their implications depending on direction and time of day. The delay analysis result was converted to a monetary value that provided insights on cost of delay and determining feasibility of a potential expansion project.
Provided firsthand perspective on traffic analysis methods, data collection process and management. Also, gained insights into statistical analysis of raw data collected from field along with preparation for simulation of the system.
Determining Condition of Road Surface and Contribution of Roadside Frictions Towards Traffic Flow to Estimate Effective Road Width
In this study, the condition survey of the corridor from Russel Square to Panthapath Intersection was conducted to identify various geometric and operational conditions of the road. The survey was conducted manually, by walking survey method with photographs. Various roadway features, availability of constituents was observed, noted, and analyzed. Causes of side-friction and loss of road width were identified with approximate loss incurred by the deterrents. Various recommendations have been made to improve the condition of the road, both geometrically and operationally.
Expand for more details regarding the research ⇣
Figure: Methodology of The Study
Abstract
A condition survey is the process of collecting data to determine the structural integrity, distresses, skid resistance, and overall riding quality of the pavement. The way surveys are conducted has vital influence on designs, on production of quantities and cost estimates and finally on execution of the work. In this study, the condition survey of the corridor from Russel Square to Panthapath Intersection was conducted to identify various geometric and operational conditions of the road. The survey was conducted manually, by walking survey method with photographs. Various roadway features, availability of constituents was observed, noted, and analyzed. Causes of side-friction and loss of road width were identified with approximate loss incurred by the deterrents. The study found the road to be adequate in terms of road conditions, however, aligning the condition with the initial purpose of the road made the condition to be unsatisfactory. Side frictions were constantly present with high frequency, leading to a low level of service at times. Various traffic control devices were installed in places, but they were dysfunctional or not operational, leading to flow being hindered. Furthermore, pedestrian crossing facilities were found to not be adequate with respect to the road being a link road with high-speed traffic. Various recommendations have been made to improve the condition of the road, both geometrically and operationally.
Figure: Example of Effective Road Width Result Summary
Outcomes
Identified congestions and pain points along busy route.
Mapped the overall corridor for potential traffic inflow, outflow, hindrances and weaving
Determined causes of side friction and corresponding reduced road width.
Road improvement measures were identified and policy recommendations were suggested.
Projects
Projects
Creating a Traffic Sign Detection and Recognition Pipeline Combining Yolov3 and Faster R-CNN
Existing literature tend to focus heavily on either the detection aspect or the classifier aspect of computer vision, as they are totally diversified and vast topics themselves. However, for driving assistance systems and/or ITS (Intelligent Transportation Systems), a pipeline is established to complete these tasks in real time. Thus, to utilize my learnt skill, to experience the complexity of such a pipeline and associated errors, bugs, I decided to combine Yolov3, a CNN(Convolutional Neural Network) based very fast object detector that suits the need of such application. Using GTSDB(German Traffic Sign Database) to train the Yolov3 model on traffic sign detection, I further incorporated additional CNN (Faster R-CNN) to recognize and classify the traffic signs in the images.
Selection of Important Features and Predicting Wine Quality
A publicly accessible dataset of the physio-chemical properties of a selection of Portuguese Vinho Verde wines was provided for me to train three models (Neural Network - Decision Tree - K-Nearest-Neighbors) to predict different boundaries of the properties of a high-quality white wine. It had the minor but important dataset issue(null, string manipulation) that was important to fix. Then the data visualization was a challenge as there were so many options to do it but finding the best one to suit this dataset. And modeling was a challenge in the sense of trying out different parameters for the best fitting model for this dataset. It was an opportunity for me to improve by learning more about how the parameters works and how more optimization possible to get better accuracy in this case.
Mechanical Properties of Recycled Aggregate Concrete for Rigid Pavements Incorporated With Rice Husk Ash & Nylon Fiber
Bangladesh aspires to establish a sustainable environment and economy; as a result, recycled aggregate concrete (RAC) is receiving more attention than ever before as a possible replacement for natural aggregate concrete in Rigid Pavements. To improve RAC's credibility, its performance in terms of strength and durability must be optimized. Rice husk ash (RHA) is a waste material, and nylon fiber (NF) is a non-biodegradable material that is commonly accessible materials in this country. Incorporating these two materials into concrete is a way of recycling and saving the environment. The research developed a unique technique for improving the mechanical behavior of recycled aggregate concrete (RAC) by the joint integration of RHA and NF. This project won the 2nd prize at American Concrete Institute (ACI) Concrete Projects Competition 2022.
Practices With Datasets
Practices With Datasets
London Bike-Sharing Demand Prediction
The dataset provided cycling data from cycling.data.tfl.gov.uk by Government of the United Kingdom and other weather related data. Exploratory data analysis was conducted, variations were analyzed, correlations and outliers were checked and dealt with, and finally, various machine learning models were trained using the dataset and r^2 values were used as the metric of comparison. Random Forest Regression was found to be performing the best, and various recommendations for future works were made.
Google Sites
Report abuse