基于深度学习的车流量分析与预测.zip

# Traffic_Analysis_and_Prediction This is a capstone project on Traffic Prediction by Team SciPy. This is our final Capstone Project in the Hamoye Data Science Summer 2022 Internship. # Introduction Traffic congestions is rising in cities around the world. Contributing factors include expanding urban populations, aging infrastructure, inefficient and uncoordinated traffic signal timing and a lack of real-time data. Given the physical and financial limitations around building additional roads, cities must use new strategies and technologies to improve traffic conditions. In this project, we will analyze the provided Traffic dataset on [Kaggle](https://www.kaggle.com/datasets/fedesoriano/traffic-prediction-dataset). The provided dataset includes hourly traffic data on four different junctions in a particular city. # Aim of the Project The aim of this project is to derive insights from this analysis and then build a model/models that can be used to predict traffic at particular times at the different junctions in the city. # Dataset Overview The traffic data that was used in this end-to-end project was sourced from [Kaggle](https://www.kaggle.com/datasets/fedesoriano/traffic-prediction-dataset) and this dataset contains 48,120 rows and 4 columns. These columns are: **DateTime:** this contains the time at which sensors collect traffic data at the different junctions. The traffic data is collected every hour. **Junction:** this represents the four different junctions from which the traffic data was collected. **Vehicles:** this represents the number of vehicles at the time the sensors capture the traffic data. **ID:** this is the unique ID of the sensors. Because this project involves a time series analysis, more columns were engineered for quality analysis, and these columns include Year, Month, Day_of_Month, Day_of_Week, Day_of_Year, Date, Time, and Seconds. # Summary of Findings - There has been an upward trend of vehicles yearly in all four junctions with junction 1 having the highest upward trend - We notice an increase in the first and third junctions around the month of June, this might be due to summer activities that happen around that time. - We notice a daily increase in Vehicular movement in all the junctions except junction four which started recording data in January 2017. - With the exception of junction 4, We notice the data increasing during the morning time, around 6 am, staying steady throughout the afternoon, and decreasing during the evening time around 8 pm. - We also notice that we have less traffic during the weekend and steady traffic during the weekdays. - Junction 4 was created to reduce the overall traffic situation on the axis which seemed to work. # Participation of Team Members for the Successful Completion of this Project Chisom Promise Nnamani - Project Lead Pearse Jim - Assistant Project Lead Victoria Udoh - Query Analyst ## Exploratory Analysis Task Group (Active Members) Chisom Promise Babatunde Raji Oguamanam Chinyere Zainab Mohammed Bernard Boateng Lakshay Arora Bissalla Daniel Subair Hussein Samuel Nnamani ## Modeling/Deployment Task Group (Active Members) Pearse Jim Gozie Ibekwe Babatunde Raji Zainab Muhammed ## Documentation/Presentation Slides Task Group (All Slides Perfectly Fit for the Presentation of the Project to Stakeholders) Victoria Udoh - Project Documentation Pragati Thakur - Presentation Slides (Main) Fidel Imaseun - Presentation Slides 2 Djardo Isaac - Prentation Slides 3 Okonkwo Bertram - Presenter Joshua Obikunle - Presenter Emekobong Udoh Odion Sonny-Egbeahie ## Dashboard Design Task Group(Active Members) Chisom Promise Omotayo Waheed # Model Deployment He’s a [link](https://team-scipy-traffic-predictor-ap.streamlitapp.com/) to the deployed model. This model is used to predict the number of vehicles that will be at a particular junction at a particular date and at a particular point in time. # Dashboard Report Here's a [link](https://bit.ly/Traffic-Data-Analysis-Dashboard-by-Team-SciPy) to the dashboard report that we created at the end of this project. In these reports, we tried to visualize the insights derived from the analysis of the traffic data like which junction had the most vehicular movement at a particular time of the day, month and year. # Presentation Slides Here's a [link](https://bit.ly/Traffic-Data-Analysis-Slide-Presentation-by-Team-SciPy) to the main presentation slides that we used used to present the process of the project and our finding to the judges and experts in the field. # Instructions For Team - The Traffic_Prediction_EDA ipynb file contains the data wrangling, analysis and visualization proceses. - Traffic Prediction Model ipynb file containing the feauture engineering and modeling processes. - Traffic Prediction Documentation contains the description of the project process from data gathering to model deployment - The traffic excel file contains the original data as it was obtained from Kaggle. - The traffic_clean excel file is the cleaned file that was used for the data visualizations and dashboard design. - The visualizations.zip contains the screenshots/photos of the visualizations used to design presentation slides. - plk files contain the trained model for each of the junctions. - trafic_predictor_app.py contains the app deployment code. - Requirements.txt file contains the packages and libraries used for the project with their listed versions. - Presentation Slides are slides that all fit for presenting and communicating the processes and insights derived from the prject to experts and stakeholders.