big data to avoid weather related flight delays ppt


Post on 23-Jan-2017




6 download

Embed Size (px)


  • Big Data To Avoid Weather Related Flight


    Pushkar Zagade

    27 March 2015

    Pushkar Zagade Tutorial Tables 1/30

  • ASeminar

    OnBig Data To Avoid Weather Related Flight Delays

    Presented ByMr.Pushkar Girish Zagade

    Roll No. [TE-49]Under the Guidance of

    Prof. Neelima SatputeDepartment Of Computer Engineering


    Academic Year 2014-15Pushkar Zagade Tutorial Tables 2/30


    I Introduction To Domain

    I Identified Problem

    I Motivation

    I Purpose And Scope

    I Objectives and planned Outcomes

    I Literature Review (Existing Methods)

    I Proposed System Components and Working Process

    I System Features

    I Mathematical Model and Analysis

    I Algorithmic Survey to finalize the algorithm

    I Survey of the selected Algorithm

    I Results

    Pushkar Zagade Tutorial Tables 3/30


    I Implementation technology

    I Implementation using data tables

    I Comparative outcome graphs

    I Merits and Demerits

    I Conclusion and Future Scope

    I References

    I Questions And Answers

    Pushkar Zagade Tutorial Tables 4/30


    I This topic is generally belongs to weather forecasting thatis how we implement Big Data computing for futureweather prediction.

    I In this lecture we will see how big data can be used toavoid weather Related Flight Delays.

    I We will see Following Concepts in this session:

    I What Is Big Data ?

    I How We Use Big Data Computing Method ForFuture Weather Prediction ?

    I Data Mining in field of Bid Data To predictweather.

    Pushkar Zagade Tutorial Tables 5/30


    I As we know because of bad weather everyday lots ofFlights has been canceled or dalayed.

    I This is a big SOCIAL ISSUES we need to solve to avoidflight delays.

    I As a Software Engineer , Engineers from University ofMichigan developed a Big Data Computing method topredict the future weather .

    I By using Big Data Computing method , they try topredict the future weather to avoid weather related flightdelay.

    Pushkar Zagade Tutorial Tables 6/30


    I Organization Need Petabyte Of Memory To Store LargeAmount Of Data.

    I The Speed Of The Buisness For The Data Is IncreasingRapidly.

    WHY THE DATA IS SO BIG?They are now creating 2.5 quintillion bytes of datadaily-Of the total data 90 percent of the data in theworld today is created in last two years.

    Pushkar Zagade Tutorial Tables 7/30


    I At present Social Media need large amount of data tostore and analyze it.

    I To give access to different users through out the world.

    I To maintain security of data.

    I For every domain , there should be large amount of datathat need to be used and analyze for new development.

    Pushkar Zagade Tutorial Tables 8/30

  • Problems With Big Data....

    I The main problem with large data is to handle such largeamount of data and it is very difficult to analyze .

    I Processing it using conventional systems takes too long.

    I It is very expansive.

    Pushkar Zagade Tutorial Tables 9/30

  • Purpose And Scope:

    Need Of Big Data For Weather Forecasting

    I Weather is everything that affect our lives on a daily basisand has the power to make permamnent changes to apersons life or buisness.

    I The Climate Corporation processes large weather datasets in excess of 50 terabytes of data.

    I On a daily basis , the corporation utilizes 2.5 millionweather measurements.

    Pushkar Zagade Tutorial Tables 10/30

  • Objectives and planned Outcomes

    I The chief goal is to enable airlines to anticipateand deal with delays before they happen, says AmyCohn, a U-M associate professor in industrial andoperations engineering who researches airlineindustry operations.

    I Big Data Management System Provides WeatherCompanies To Get Previous Decades WeatherTrends And Patterns.

    I By Analysing These Large Data Sets, WeatherPredictions Are Becoming More Accurate AndBeing Quicker.

    Pushkar Zagade Tutorial Tables 11/30

  • Existing Method

    I Prediction of Weather through Satellite is mostcommonly used method of weather prediction.

    I Since it give appropriate result , but it predict the weatherof some limited days about 2 weeks.

    I Big Data Computing method overcome this limitation ofSatellite system , So future weather become easy topredict.

    Pushkar Zagade Tutorial Tables 12/30

  • Brief WorkingResearch In The University Of Michigan

    I The students from the University of Michigan havestarted a new research which helps in understanding theweather of a particular place.

    I Theyve gathered more than 10 years of hour-by-hourweather observations and domestic fight data, and areusing advanced data analytics to spot patterns and helpairlines manage more efficiently..

    I The project uses public data that has been available foryears, its size and scope make it unique, says BrianLemay, a U-M doctoral student in industrial andoperations engineering who leads the project.

    Pushkar Zagade Tutorial Tables 13/30

  • Brief WorkingResearch In The University Of Michigan

    I Were the first to gather this data in one place andapply this level of computing power to it, Lemay said.That enables us to do a very sophisticated analysis ofhow weather and flight delays are connected and go farbeyond individual airports.

    I The chief goal is to enable airlines to anticipate and dealwith delays before they happen, says Amy Cohn, a U-Massociate professor in industrial and operationsengineering who researches airline industry operations.

    I This is a very creative and new process. It could lead tounderstanding similarities in the weather in the past years.It could be of help in predicting the weather in thefuture.With the help of this data, the flights can becautious of bad weather in advance.

    Pushkar Zagade Tutorial Tables 14/30

  • Application

    Imagine youre scheduled to fly out of Detroit four hoursfrom now and theres a storm in Atlanta. Cohn said.The airline could use this data to determine that thestorm in Atlanta is likely to delay your plane. They couldthen contact you and offer you a seat on an alternateflight. You save time, and the airline doesnt have toaccommodate you on a later flight after the delayhappens.

    Pushkar Zagade Tutorial Tables 15/30

  • Data Mining In Big Data

    I Generally, data mining is the process of analyzing datafrom different perspectives and summarizing it into usefulinformation - information that can be used to increaserevenue, cuts costs, or both.

    I Data mining software is one of a number of analyticaltools for analyzing data. It allows users to analyze datafrom many different dimensions or angles, categorize it,and summarize the relationships identified.

    I Technically, data mining is the process of findingcorrelations or patterns among dozens of fields in largerelational databases.

    Pushkar Zagade Tutorial Tables 16/30

  • System Features

    I Large quantities of dataThe volume of data so great it has to be analyzed byautomated techniques e.g. satellite information, creditcard transactions etc.

    I Complete dataThis data is the characteristic of all data collection.

    I Easy data structurePrevious data are available which are complete so itbecome easy to analyse it.

    Pushkar Zagade Tutorial Tables 17/30

  • Classification And Prediction Algrithm

    I The previous data should be collected in order to analysedit.

    I This algorithm consist of removal of unwanted data andselection of proper data.

    I Apply Back Propagation Algorithm and classify data intocollection and future prediction.

    I About 80 percentage should be accurate by using thistechniques.

    Pushkar Zagade Tutorial Tables 18/30

  • Data Flow Diagram

    Pushkar Zagade Tutorial Tables 19/30

  • Survey Of Back Propagation Algorithm

    I In order to perform a Back Propagation Algorithm aprogram or logic has to be created. What will be thechange on other parameters by changing any oneparameter, will be observed.

    I After predicting that, what will be the weather in futureafter some period, the Classification method will takeplace.

    I In Classification method, it will display what will be thefuture weather, whether it will be sunny day, rainy orcloudy day what will be the change in wind speed,humidity etc.

    I The Classification Technique will help for taking someprevention from the climatic hazards.

    Pushkar Zagade Tutorial Tables 20/30

  • Mathematical Model

    Where, A -.Collection Of DataB - Waste DataC - Useful DataD - Apply BackPropagation Algo.E - PredictionF - Collection Of InformationG - Classification

    Pushkar Zagade Tutorial Tables 21/30

  • Implementation Of Technoloy


    I Data Collection

    I Pre-processing

    I Data Transfer

    I Data Mining

    I Prediction of Future Weather using ANN by BackPropagation Algorithm

    I Classification

    Pushkar Zagade Tutorial Tables 22/30

  • Data Tables

    Pushkar Zagade Tutorial Tables 23/30

  • Comparative Outcome Graph

    Pushkar Zagade Tutorial Tables 24/30

  • Merits

    I Predict future trends, customer purchase habits

    I Help with decision making

    I Improve company revenue and lower costs

    I Market basket analysis

    I Fraud detection

    Pushkar Zagade Tutorial Tables 25/30

  • De-Merits


View more >