Categoría: SmartCities

Sprint 3 Review

This sprint we did the following:

  • Rebuilt the backend response in different endpoints.
  • Redeployed all the infrastructure, because the services were down.
  • Changed the credentials needed for the Redis.
  • Developed the frontend needed to show in real time the information (bikes available and the ones disabled).
  • Developed a serverless function which retrieves the data on real time and calculates a new output every time we want to consult the web page.

week 6 planning

This week we will do the following:

  • Make a PoC using Random Forest
    • Develop in Python a Markov Chain with the datasets disposed
  • Test backend service
    • Do unit testing
    • Do integration testing
  • Complete the Homepage view (frontend)
    • Only boilerplate
    • UX is the priority
  • Configure CircleCI
    • For automating tests
    • In a short term will help to deploy
  • Configure TCP sockets in backend
    • Impment TCP Connection Address in GO

Week 5 planning

This week we will do the following:

  • Research using Data IQ
    • Analyze the distribution of the data
  • Make a PoC using Random Forest
    • Develop in Python a Markov Chain with the datasets disposed
  • Define endpoints needed for backend
    • Make a PoC for every endpoint using Redis
    • Testcases included
    • Try to do a TDD strategy
  • Complete the Homepage view (frontend)
    • Only boilerplate
    • UX is the priority
  • Configure CircleCI
    • For automating tests
    • In a short term will help to deploy
  • Configure TCP sockets in backend
    • Impment TCP Connection Address in GO

Week 4 Review

This week so far, we did and delivered the following:

  • Backlog with detailed tasks for every sprint planning:
  • As well, we have redesigned the problem:
    • Now we will analyze clusters of respective regions, neighborhoods in every city which must have MiBici.
    • The new model which we are going to implement will be a Bayesian Classification/Markov Chains instead of the Multinomial Logistic Regression.
    • We are analyzing several solutions in Python, one that we have considered so far is the Random Forest algorithm.
  • We have defined the team organizationScreen Shot 2019-02-10 at 21.44.43

Week 3 review

  • Machine Learning Model
    • To solve the inital problem (which is to predict a range of bikes available or unavailable per station), we are planning to use a Multinomial Logistic Regression Model.
    • We have defined that there are 4 important variables:
      • Initial date of trip (which can be split into year, month, day and hour)
      • Final date of trip (which can be split into year, month, day and hour)
      • Departure (id).
      • Arrival (id).
  • Matlab – We did a PoC using Matlab about comparing the initial time and final time of each user’s trip.
    • Screen Shot 2019-02-02 at 17.31.20.png
  • Domain requirements
    • Currently in progress.
  • Proposal Technologies
    • Go
    • React/Redux.
    • Google Cloud/App Engine
    • Docker
    • Matlab
    • Python/SUMO
  • Bakckend architecture
    • Currently in progress
  • MockupsScreen Shot 2019-02-02 at 17.38.40.png
  • Project Charter
  • PROJECT CHARTER
    GENERAL PROJECT INFORMATION
    PROJECT NAME MiBici Battery Swap
    PROJECT SPONSOR Bosch
    PROJECT MANAGER Alberto Briseno
    EMAIL ADDRESS Jose.Briseno@mx.bosch.com
    PHONE NUMBER 52 33 1579-1263
    ORGANIZATIONAL UNIT Bosch ML Team
    PROCESS IMPACTED public transport
    EXPECTED START DATE January 31, 2019
    EXPECTED COMPLETION DATE May 1, 2019
    EXPECTED SAVINGS $0
    ESTIMATED COSTS $0
    GREEN BELTS ASSIGNED NA
    BLACK BELTS ASSIGNED NA
    DESCRIBE THE PROBLEM OR ISSUE, GOALS, OBJECTIVES, AND DELIVERABLES OF THIS PROJECT
    PROBLEM OR ISSUE MiBici is a project in where you pay a membership to use one bike from the different stations around the city, after you pick one, the bike must be returned to one of the stations already mentioned. The project is about predicting the times and in which zones are used the bikes, in order to deliver a simulation associated with the data analyzed.
    PROJECT GOAL The bikes are intended to be replaced in a short period of time for electric ones. The team involved in this project required to know the information that we must deliver, in order for swapping the batteries.
    BUSINESS CASE Public transport/ Encourage a healthy and better lifestyle.
    GOALS / METRICS Machine Learning Predictive Model/ Bikes simulator in a real time traffic approach.
    EXPECTED DELIVERABLES Web platform with the stations, the number of bikes in each station and in estimated range of time.
    DEFINE THE PROJECT SCOPE AND SCHEDULE
    WITHIN SCOPE Machine Learning Predictive Model/ Bikes simulator
    OUTSIDE OF SCOPE Web platform
    TENTATIVE SCHEDULE KEY MILESTONE START COMPLETE
    Form Project Team / Preliminary Review / Scope 01/25/19 01/31/19
    Finalize Project Plan / Charter / Kick Off 02/05/19 11/02/19
    Development Phase 02/12/19 04/09/19
    Measurement Phase 02/26/19 04/09/19
    Analysis Phase 03/05/19 04/16/19
    Improvement Phase 03/08/19 04/20/19
    Control Phase 03/19/19 04/25/19
    Project Summary Report and Close Out 04/30/19 05/03/19
    DEFINE THE PROJECT RESOURCES AND COSTS
    PROJECT PARTICIPANTS Cesar, Constanza, Katia, Mario, Victor, Mark, Montserrat, Gabriel, Alberto, Julia, Samuel, David, Estefania and team from Bosch
    SUPPORT RESOURCES
    Matlab, SUMO, accesability to MiBici data and web frameworks.
    SPECIAL NEEDS
    Computers that can support the simulations.
    COST TYPE VENDOR / LABOR NAMES RATE QTY AMOUNT
    LABOR Planning 2 wk $0
    LABOR Developing 2 wk $0
    LABOR Testing 3 wk $0
    LABOR Developing 2 wk $0
    LABOR Testing 3 wk $0
    LABOR Deployment 2 wk $0
    TOTAL COSTS $0
    DEFINE THE PROJECT BENEFITS AND CUSTOMERS
    PROCESS OWNER Alberto Briseno
    KEY STAKEHOLDERS Cesar, Constanza, Katia, Mario, Victor, Mark, Montserrat, Gabriel, Alberto, Julia, Samuel, Estefania, Ismael, David, team from Bosch and people who use public bikes.
    FINAL CUSTOMER
    People which has the control management over MiBici logistic.
    EXPECTED BENEFITS People will be able to have bikes available at all stations with charged batteries (the bikes are ment to be electric).
    TYPE OF BENEFIT DESCRIBE BASIS OF ESTIMATE EST BENEFIT
    SPECIFIC COST SAVINGS NA $0
    ENHANCED REVENUES NA $0
    HIGHER PRODUCTIVITY (SOFT) NA $0
    IMPROVED COMPLIANCE NA $0
    BETTER DECISION MAKING NA $0
    LESS MAINTENANCE NA $0
    OTHER COSTS AVOIDED NA $0
    $0
    DESCRIBE PROJECT RISKS, CONSTRAINTS, AND ASSUMPTIONS
    RISKS We are expecting to learn technolgies which we haven’t used and deliver the best as we can the Machine Learning and simulation part.
    CONSTRAINTS Time, project meetings
    ASSUMPTIONS The stakeholders will participate actively and the user stories, engineering tasks and requirements may be well stablished and may not change very constantly.
    Prepared by: Kassem’s Team Date: January 29, 2019