Scribd Logo
Upload

Welcome to Scribd!

  • 23 views

    Project Presentation

    Uploaded by

    Manisha Shinde
    The document discusses predicting food delivery times using machine learning models. It describes preparing a dataset containing delivery details, exploring the data, building models using XGBoost and hyperparameter tuning, and achieving a RMSE of 7.28 minutes on delivery time prediction. Key factors affecting time were identified as delivery person ratings and distance.

    Copyright:

    © All Rights Reserved
    Download as PPTX, PDF, TXT or read online from Scribd

    Project Presentation

    Uploaded by

    Manisha Shinde
    23 views14 pages
    The document discusses predicting food delivery times using machine learning models. It describes preparing a dataset containing delivery details, exploring the data, building models using XGBoost and hyperparameter tuning, and achieving a RMSE of 7.28 minutes on delivery time prediction. Key factors affecting time were identified as delivery person ratings and distance.

    Original Title

    project presentation

    Copyright

    © © All Rights Reserved

    Available Formats

    PPTX, PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    The document discusses predicting food delivery times using machine learning models. It describes preparing a dataset containing delivery details, exploring the data, building models using XGBoost and hyperparameter tuning, and achieving a RMSE of 7.28 minutes on delivery time prediction. Key factors affecting time were identified as delivery person ratings and distance.

    Copyright:

    © All Rights Reserved
    Download as PPTX, PDF, TXT or read online from Scribd
    Download as pptx, pdf, or txt
    23 views14 pages

    Project Presentation

    Uploaded by

    Manisha Shinde
    The document discusses predicting food delivery times using machine learning models. It describes preparing a dataset containing delivery details, exploring the data, building models using XGBoost and hyperparameter tuning, and achieving a RMSE of 7.28 minutes on delivery time prediction. Key factors affecting time were identified as delivery person ratings and distance.

    Copyright:

    © All Rights Reserved
    Download as PPTX, PDF, TXT or read online from Scribd
    Download as pptx, pdf, or txt
    of 14

    Food Delivery Time

    Prediction

     PRESENTED BY: 1.Deepali Ghumare


    2.Manisha Shinde
    3.Pallavi Kale
    4.Akash Satapute
    5.Vaibhav Hegadkar
    Introduction
    Food delivery is a service that allows customers to order food from restaurants and
    have it delivered to their doorstep.

    More businesses are moving online these days, and consumers are ordering online
    instead of traveling to the store to buy. Zomato and Swiggy are popular online
    platforms for ordering food products. Other examples are Uber Eats, Food Panda, and
    Deliveroo, which also have similar services. They provide food delivery options.

    If the order is complete, a partner will pick up and deliver the meal to the given
    address via a delivery service. In online food-ordering businesses, delivery time is
    critical. As a result, estimated food delivery time prediction to reach the buyer’s
    location is critical.

    Objectives
    1. What factors can affect the time in delivering food from the restaurant to the destination location?
    2. How much food delivery time prediction accuracy performance?
    Data Preparation
    General Info

    • Dataset consists 11 3 9 7 rows and 20 columns. Then the


    dataset also consists of 5 numerical data and 1 5 categorical
    data
    • The problem faced is a regression problem, namely
    predicting the time needed to deliver food (Time taken)
    • There are no missing value, duplicated value, and odd data in this
    dataset
    • All numerical values contained in the dataset are quite
    reasonable
    geodesic function
    geodesic function

    Data Preparation
    Distance Calculation
    To get the time needed to deliver food, the distance between
    the restaurant and the delivery location is needed. To get the
    required distance, we use geodesic function which can be
    used to calculate the distance between 2 locations by
    utilizing longitudes and latitudes.

    Geodesic Function
    Library- geopy.distance.geodesic
    Exploratory Data Analysis
    Univariate Analysis

    Outliers Check Data Distribution Check


    Extreme outliers at Delivery_person_Ratings column are on lower Delivery_person_Age and Time_taken(min)
    boundary (< 3.9), meanwhile distance column have extreme column have relatively normal data distribution
    outliers in upper boundary (> 31.9) Delivery_person_Ratings has negative skew data
    distribution, meanwhile distance column has
    positive skew data distribution
    Exploratory Data Analysis
    Bivariate Analysis

    There is consistent relationship There is a linear relationship between the time taken There is an inverse linear relationship
    between the time taken and the to deliver the food and the age of the person who between the time taken to deliver the food
    distance travelled to deliver the delivering the food. It looks like person with the young and the delivery person ratings. It looks
    food. It looks like majority food age able to take less time than person with old age to like person with the higher ratings take
    delivered within 25-27 minutes deliver the food to customers a less time to deliver the food than person
    regardless of distance with low ratings
    Exploratory Data Analysis
    Bivariate Analysis

    It looks like there is not much difference between the time taken depending on the vehicle they are driving and the type of food
    they are delivering
    Data Pipeline Architecture
    1. AWS Instance Creation:

    • Provision an Amazon EC2 instance to serve as the infrastructure for our data pipeline.

    2. Zookeeper and Kafka Setup:

    • Install and configure Apache Zookeeper for distributed coordination.


    • Install and configure Apache Kafka for real-time data streaming.

    3. S3 Bucket Creation:

    • Create an Amazon S3 bucket to store and manage data.


    4.Python Data Loading:

    • Develop a Python script to load data into the S3 bucket.


    • Utilize the Boto3 library to interact with AWS services programmatically.
    5.Athena Data Querying:

    • Set up an Amazon Athena database to query data stored in the S3 bucket using SQL-like queries.
    • Define tables and partitions to optimize query performance.
    6.Power BI Connection through ODBC:

    • Install and configure the Athena ODBC driver on the EC2 instance.
    • Set up an ODBC data source to connect Power BI to Amazon Athena.
    Feature Engineering

    Initial Feature Final Dataset


    Dataset Engineering (Train & Test
    Data)
    Unused Feature Drop
    Outliers Handling
    One Hot Encoding
    Train Test Data Split
    (70:30)
    Modeling a n d Evaluation
    Algorithms Training
    Based on each R M S E (Root Squared Mean Error) result
    from
    5 trained algorithms, best algorithm is XGBRegressor
    because have the smallest error than other algorithms.
    Performance of the best model will try to be improved
    with hyperparameter tuning

    R M S E (Root Squared Mean Error) Formula


    Modeling a n d Evaluation
    Hyperparameter Tuning

    Parameters to be tuned

    Parameter tune using GridSearch CV

    Best parameter

    Model train using best parameter

    Final model performance


    Modeling and Evaluation
    Feature Importances
    Delivery_person_Ratings variable become the most
    importance feature in this model, followed by Distance and
    Delivery_person_age variables. Then, we can see
    Delivery_person_Ratings variable has a negative
    contribution when its values are high, and a positive
    contribution on low values

    Prediction on N e w Data
    Conclusion and Recommendation
    Conclusion
    1. Rating of person in previous delivers become is the most influential factor on the delivery time of food to the
    destination location. Person with the higher ratings take a less time to deliver the food than person with low ratings
    2. Model has R M S E score 7.28 and that means error between delivery time prediction and delivery time actual is 7.28 minutes

    Recommendation
    The rating obtained by the deliveryman is a representation of the deliveryman's performance in delivering food to the intended
    location in terms of delivery time. Of course this is a potential loss of customers if this continues to happen. Delivery time
    performance needs to be maintained so that the rating obtained is high and customer trust can still be maintained.
    The R M S E value of the model can be used as a guarantee of delivery time performance which can be given to the customer so that
    as much as possible the delivery time is not more than the existing R M S E score (delay in delivery time of not more than 7.28
    minutes).

    Documentation : Github Repository


    THANK
    YOU
    +62 8 7 8 7375 6695

    [email protected]

    fooddeliveryx456

    You might also like