Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Smart Suggestions - Part I

Smart Suggestions - Part I

Sricharan

March 13, 2016
Tweet

More Decks by Sricharan

Other Decks in Education

Transcript

  1. Driver navigation systems for last-mile
    shipping efficiency
    Mid-term Presentation | Sricharan Chiruvolu | BL.EN.U4CSE12505

    View Slide

  2. Company Overview
    2

    View Slide

  3. • Hyper-local logistics service.
    • One of the largest fleet of delivery drivers.
    • Merchant and driver platforms.
    • B2B. Food, Groceries…11 others.
    • 25000 orders a day.
    • McDonalds, Pizza Hut, KFC…4000 others.
    • Tech Stack: Rails, MySql, MongoDb, Redis,
    Angular, Android, Scala, NodeJS and Django.
    3

    View Slide

  4. Tasks and Projects
    4

    View Slide

  5. Tasks and Projects
    1. Driver reassignment by merchants for a live order.
    2. Driver navigation to last-mile customer location.
    3. Smart suggest for drivers.
    4. Order stockout analytics fixes. (End-term presentation)
    5. Cross locality preferences analytics. (End-term presentation)
    5

    View Slide

  6. Driver Reassignment
    6

    View Slide

  7. Driver Reassignment
    • Add an option for the merchant to reassign the driver for
    an order during cancellation.
    • Decrease the number of cancellations.
    • Assign to the next nearest free driver.
    • Reasons for reassignment:
    1. Delivery driver refused to come.
    2. Driver didn't come for a long time.
    • Reassignment workflow.
    7

    View Slide

  8. start
    stop
    Reassign Order?
    active_leg?
    &&
    next_driver_available?
    &&
    correct_cancellation_reason?
    Cancel Order
    Order reassign to next free driver
    active_leg?
    Cancel Order?
    Error Message
    No No
    Yes Yes

    View Slide

  9. Driver Navigation to Customer
    Location
    9

    View Slide

  10. Driver Navigation
    • Last mile navigation.
    • Capture customer location.
    • Send customer location for a live trip.
    • MQTT(MQ Telemetry Transport) and GCM (Google
    Cloud Messaging) protocols.
    10

    View Slide

  11. Last-mile navigation
    11

    View Slide

  12. MQTT activity at Runnr
    12

    View Slide

  13. Smart suggest
    13

    View Slide

  14. Smart suggestions
    • Demand - supply gaps in localities.
    • Demand prediction and supply input.
    • Calculating demand and supply percentile.
    • Surge pricing prerequisite.
    • Driver application and operations dashboard
    views.
    14

    View Slide

  15. Smart suggest on operations dashboard
    15

    View Slide

  16. Smart suggest on driver application
    16

    View Slide

  17. Smart suggest on driver application (offline flow)
    17

    View Slide

  18. Next…
    • Stockout data fixes:
    • Find stockout orders and predict stockout
    quotient.
    • Cross-locality orders:
    • Cost prediction (machine learning)
    18

    View Slide

  19. Thank you.
    19
    NAME
    SRICHARAN
    (BL.EN.U4CSE12505)
    EXT. GUIDE
    Mr. MUKUNDA & Mr. MOHIT
    AGARWAL
    INT. GUIDE Ms. SANGITA KHARE

    View Slide