I declare that this report entitled "OPPORTUNISTIC RIDE-SHARING USING GOOGLE MAPS ANALYTICS" is my own work, except as cited in the references. This is an automated ride-sharing model that can improve the efficiency of matchmaking of rides that populistically share similar itineraries and schedules. The GPS navigation in the smartphone is used as an indicator that can track the location visited by the user.
User requires to turn on the GPS location services and mobile data in the smartphone and the system will be able to start tracking and capturing the visited location and departure date and time, the collected data will be uploaded to the Firebase database. In traditional ridesharing or messaging, drivers within a predetermined area of the Internet are assigned to riders based on proximity distance. This automatic ride sharing system is a user friendly system that the user only had to open the system once and the system will continue to track the location you visit.
LIST OF TABLES
LIST OF ABBREVIATIONS
Introduction
- Problem Statement
- Motivation
- Project Objective
The problem with the existing ride-sharing services is the time-consuming nature of matching drivers with users. Ride sharing is one of the most beneficial actions to reduce urban traffic. Therefore, the ride-sharing services are growing rapidly in the market, here are some commercial on-demand ride sharing services such as Grab, Uber and Lyft.
However, the opportunistic opportunities of ride-sharing services still exist in inefficient performance when the matching process for the driver and the user. As we mention the scenarios above, the current ride sharing services are not effective, efficient and time consuming. This automated ride sharing system can match the driver and user who have similar routes and schedules and reduce the problem of time consumption.
Literature Review 2.1 Review of existing works
- Comparison. Of Existing Works with Proposed System
Chowdhury et al. 2016) proposed a campus ride sharing platform that implemented the Proximity Search panel, Trip List (Circle Priority), Rating Option panel, and Notification panel. Farin et al. 2016) proposed a dynamic framework for vehicle pooling and ride sharing that provided security, accessibility and user identification. In the previous case, they proposed a two-stage approach that is more effective in the practice of trip planning.
However, the setup of the vHike system via Bluetooth Chat still requires the user to find and launch a card with the driver to confirm the ride, after which the system is an inflexibility search technique in ride sharing. While the Campus Ride Sharing Platform also has low accuracy to locate a rider's driver because this system used the postal method to find a driver for a ride. For our proposed system, opportunistic ride sharing using Google Maps Analytics can outperform other systems in terms of tracking accuracy, flexibility, and time-consuming to search and match a ride share.
System Design 3.1 Overview of System Design
- Data Collection Program
- User Registration
- Location Data Collection
- Data Labeling
- Reference of data
- Model Training
The figure 3.3.2.1 shows the interface of the main screen of the location tracker, and the GPS is in tracking mode. From the figure 3.3.2.3 shows that the location manager is used for the object that starts and stops the delivery of the location related events in the location tracker. Based on the figure 3.3.2.6 above shows that location data that stores in the Firebase Realtime Database.
For the time classification part, figure 3.4.1.1 defines the time series that classifies all hours and minutes, the from. Based on Figure 3.4.1.5 showing that this is the first part of the main program, the three arguments will be declared, namely node, uid as the user ID and csv_file_path as the output file name. Based on Figure 3.4.2.1, the second part of the main data program is shown.
From figure 3.5.1 it shows that the data set will read and determine a basis to make the random number predictable and shuffle the data. Based on Figure 3.5.7, a function is defined to be used for the call to start to train a model, KNN will be selected as the algorithm to be used in the training job. From figure 3.5.4 it shows that feature shape is 2, then feature_dim in all model training is using 2.
Finally, the task will start routing after the function call in Figure 3.5.7 and the model will store the S3 bucket. Figure 3.5.10 defines a function to set the content type, serializer and deserializer of the model, the content type of the model is in text/csv, the serializer is csv_serializer and the deserializer is. Based on Figure 3.5.12, it can be seen that the model will be tested to calculate the accuracy of the model.
Based on Figure 3.6.1, which shows the entire prediction process flow, the first step is to get the endpoint as a real-time predictor, and then set the predictor details. Figure 3.6.5 shows that 3 user datasets will read and define the seed to make the random number predictable and shuffle the data. Based on Figure 3.6.6, it can be seen that the model of each of the three users will be tested to calculate the accuracy of the model.
System Methodology
- Methodologies and General Work Procedures
- Tools to use Xcode
- User Requirements
Xcode is an integrated development environment for macOS that contains a set of development tools that will be used to develop the main program of this project. The main program will be developed using Swift and is designed to allow developers to build apps for iOS, Mac, Apple Watch and Apple TV. In this project, a mobile application will be developed and the main software functions such as user authentication, extraction of 2D location coordinates from the map, location data collection and reverse geocoding will be developed in Xcode IDE.
In addition, the program will also be used to call the API key to connect to the database, which is the real-time database of Firebase, to update and retrieve user information from the database . Pycharm is an integrated development environment used in the computer programming tool, which will also be used to develop the program of this project. This program will be developed using Python, it is an object oriented programming language, it is supporting all kinds of different data formats and can use comma separated value document or JSON sourced from the web.
So the algorithms in this project are JSON data extraction and conversion data from JSON to CSV format. Using SageMaker's algorithms, this project uses the k-neest neighbors (KNN) algorithm to analyze and train the collected data. After the training job is complete, the model is hosted in the SageMaker, ready to use for predictions.
The most important data will be saved in the history, and within the history it will contain the visited location and the date time will be the key of each visited location. This is because the whole project needs GPS location services to track the user's location all the time, if the service is not turned on, it will not be able to track the user's location. Due to the type of database used in the project, mobile data is needed to load and save data to the database when data is collected.
Each time the user logs into the mobile app, the app will provide an access token that will be used to identify the user.
Implementation and Testing 5.1 User Registration
- Location Data Collection
- Model Training
- Prediction
- Test Parameter
- Scenario 1: No Driver is suggested
- Scenario 2: All drivers are going to the same destination
- Scenario 3: Only one driver going to the same destination
- Summary of Test Result
- Discussion of Result
- Project Review and Conclusion
- Limitation
If the condition is only one user goes to the same place, then the user is proposed directly. In the testing part, the tester will select the various inputs to the system and determine the accuracy of the models, algorithms and predictions. In the first scenario, between the three drivers do not have the same destination point with the user.
In another scenario, three drivers go to the same destination point with the user, but only one driver will be suggested to the user, ranked by accuracy. Test case Test description Theoretical result Actual result 2 All drivers work the same. In the third scenario, only one driver goes to the same destination as the user.
In this case, only driver 3 goes to the same destination as the user, so driver 1 will be the suggested driver. Test Case Description Test Theoretical result Actual result 3 Only the driver goes to the same thing. In the table 5-6-5-1 shows the summary of the test result, and the system had passed all test cases.
Table 5-6-5-1 shows the result obtained from the test cases, the accuracy of the system is 100 percent, which means that the system passed all the test cases. For the first test case, the system predicted that none of the drivers go to the same destination as requested by the user, and the prediction result is exactly the same as the expected result, so the result of test case 1 is valid. Then the test case is when all the drivers go to the same destination requested by the user, then there will be a suitable classification and the driver will be classified according to the accuracy of the model, and the driver with the highest accuracy will be suggested to the user.
Finally, in test case 3, if only one driver goes to the same destination with the user request, the only driver will suggest the user. In this condition, the actual result is the same as the expected result, so the result of test case 3 is valid. If the data is not enough, the model's accuracy becomes lower, which can affect the prediction results. A functional website or mobile application can be developed to use the system's algorithm.
UNIVERSITI TUNKU ADBUL RAHMAN
Introduction
Description and Methodology
Result
Objectives
The required originality parameters and limitations approved by UTAR are as follows:. i) the total similarity index is 20% or less, and. ii) Matching of individual cited sources must be less than 3% each and (iii) Matching of texts in a continuous block must not exceed 8 words. Note: Parameters (i) – (ii) exclude citations, bibliography and text matches of less than 8 words. Note The supervisor/candidate(s) must provide the Faculty/Institute with an electronic copy of the complete originality report set.
Based on the above results, I declare that I am satisfied with the originality of the final year project report submitted by my students as mentioned above. Form Title: Supervisor Comments on Originality Report Generated by Turnitin for Final Year Project Report Submission (for Undergraduate Programs).
UNIVERSITI TUNKU ABDUL RAHMAN FACULTY OF INFORMATION & COMMUNICATION
TECHNOLOGY (KAMPAR CAMPUS)
