Click here to load reader

Software Project Management for 'Weather Forecasting using Data mining

Embed Size (px)

Citation preview

Slide 1

Weather Forecasting Using Data MiningFinal Project 05/04/2016SWEN 5230-03Software Project Management1

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

ContentsSr. No.ArtifactSlide Number01Business case0302Vision4603Architecture6304Work Breakdown structure8105Software Development Plan9906Summary

2

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Weather Forecasting using Data MiningBusiness Case03/30/2016SWEN 5230-03Software Project Management3

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

OverviewBackgroundProject overviewArchitectureCost ScheduleFinancial benefit4

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

BackgroundThe ProblemWeather forecasting is a vital application in meteorology and has been one of the most scientifically and technologically challenging problem around the world.Primary users for the system are general public, aviation, fire and marine.Aviation forecasters use them in order to keep an eye on surface observations for wind shear, restrictions to visibility that could affect takeoffs and landings.[4]

5

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

BackgroundForecasters support fire weather programs by checking for relative humidity because it can have critical impact on behaviors of fire. Every member of population uses weather data on regular basis, thinking of how weather can affect your travel, activity and business decision the list of users become longer[4]Similar service providers are National Weather Services(NWS), popular sites like Weather Underground, forecast.io, Weather spark and Google6

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Project OverviewIn data mining we use analysis tools to discover patterns and relationships in data, that maybe used to make valid predictionIn this proposed software we investigate use of data mining in forecasting maximum temperature, rainfall, evaporation and wind speed.

7

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

7

Project team basedProject ManagerArchitectBusiness AnalystDesigner TesterRequirement specialist8

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

What we plan to doIn this proposed software, we plan to use the data mining techniques in forecasting weather.This can be carried out with help of Artificial Neural Network, Decision Tree Algorithm and meteorological data collected in specific time.We plan to follow Royces project lifecycle, and the software development plan and work breakdown structure explains the activities we have scheduled.Primarily, the elaboration phase will focus on getting the prototype approved from stakeholders.9

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

High Level ArchitectureFor weather prediction software, we plan to use both algorithm using both Artificial Neural Network(ANN) and Decision Trees(DT).The data mining technique which is going to be used will require historical data which is done under Data Collection component of this process.For data sets we will be using current weather services APIs like Weather.com, forecast.io, Accuweather.10

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

High Level Architecture

Image Source: Click here [1]11

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Data mining Architecture

Image Source: Click here [2] 12

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Major ComponentsData Collection: As explained earlier we would be feeding historical data to the system, this could be from specific regionData Cleaning: Under this component the data glitches like the missing data, duplicated data is found and bad data is weed outData selection: Under this stage, relevant data related to analysis is retrieved and classified under 10 attributes as shown in table below13

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Attributes of data set[Image Source [2]]14

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Major ComponentsData Transformation: Under this stage the selected data will be transformed to accepted form for data mining(Commas Separated Values)Data mining : Under this stage, the algorithms are used to analyze meteorological datasets which further results in interesting patterns which are then studied15

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

COTSHistorical climate data from various sources for specific region which can be bought from meteorological centers or airport stations

16

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Estimating CostWe shall be estimating cost using use case points, estimating use case pointsSystem Requirements:Enumerated Functional Requirements:The system should allow administrator to add historical weather dataThe system should be able to recognize patterns in temperature, humidity and wind with use of historical data

17

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Estimating CostSystem should periodically apply prediction algorithms or models on obtained data and store results to central databaseSystem shall obtain and display confidence value for each prediction given to userSystem shall support registering new user by providing email, so system can send daily weather updates to registered usersSystem shall allow users to check weather for one week ahead18

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Estimating CostSystem should be able to alert its registered users by sending storm warnings in emergencies

Non-functional requirements:System shall allow for users to get prediction for weather within almost two mouse clicksSystem should ensure that features that do not require a user to be logged in are not hidden to unregistered users19

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Estimating CostSystem should be able to run with core functionality intact from Smartphone or tabletSystem should be able to show interactive animations to users regarding current and future climatic conditionsSystem should allow language preference to usersSystem should provide graphs and maps of climate conditions

20

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Estimating CostActors and User goals:

Registered UserVisitorDatabaseHistorical data providerGraph plotterAdministrator21

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Estimating CostsUse cases:

For Administrator:Report(Feed) historical weather dataReport status Remove data Logout

22

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Estimating Costs2. For Registered user:Track weatherSearch for future predictionsGet hourly/timely weather updates in email

3. For Graph plotter:Plot sketches of weather predictions on map23

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Estimating Costs4. Visitor:Check weather statusRegister for timely updates24

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Estimating CostsEffort Estimation:Unadjusted Actor Weight Standards:

Actor TypeExampleWeightSimpleAnother system through an API1AverageAnother system through a protocolA person through a text based user interface2ComplexA person through a graphical user interface3

25

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Estimating Costs

ActorDescriptionComplexityWeightRegistered UserA registered userComplex3VisitorA unregistered userComplex3DatabaseRecords of weather information, user data and system dataAverage2Historical weather data providerProvides historical data from different sourceAverage2Graph plotterProvides visuals from raw dataSimple1AdministratorSpecial case user that updates website Complex3

26

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Estimating CostsUAW = 3 x Complex + 2 x Average + 1 x Simple = 3 x 3 + 2 x 2 + 1 x 1 = 9 + 4 + 1 = 14

Unadjusted use case weight (UUCW): In Kerners original formula, each use case is assigned a number of points based on number of transactions within the use case 27

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Estimating CostsUnadjusted Use case weights standards:

Use case complexityNumber of transactionsWeightsSimple3 or fewer5Average4 to 710ComplexMore than 715

28

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Estimating CostsUse caseComplexityWeightReport Historical weather dataComplex15Report statusComplex15Remove DataAverage10Log outSimple5Track weatherAverage10Search for future climate predictionComplex15Get timely updates in emailAverage10Plot sketches of weather predictionComplex15

29

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Estimating CostsUse CaseComplexityWeightsCheck weather statusSimple5Register for timely updatesAverage10

UUCW = 4 x Complex + 4 x Average + 2 x SimpleUUCW = 4 x 15 + 4 x 10 + 2 x 5UUCW = 60 + 40 + 10 UUCW = 11030

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Adjustment for Technical Complexity:

FactorComplexityWeightCalculated FactorDistributed System236Performance objectives236End user efficiency122Complex processing144Reusable Code122Easy to install0.531.5Easy to use0.552.5Portable224Easy to change111

31

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Estimating CostAdjustment to technical complexityFactorComplexityWeightCalculatedConcurrent use144Security155Access for third parties111Training needs122

TCF = 0.6 + (0.01 x 41) = 1.0132

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Environmental Complexity FactorFactorWeightAssessment ImpactE1Familiarity with development process1.534.5E2Application Experience0.542E3Object oriented experience144E4Lead analyst capability0.542E5Motivation155E6Stable requirements212E7Part time staff-100E8Difficult programming language-12-2

33

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Environmental Complexity FactorECF = 1.4 + (-0.03 x EFactor) = 1.4 + (-0.03 x 17.5) = 0.875

34

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Estimate Cost and ScheduleUnadjusted use case points (UUCP) = UAW + UUCW= 110 + 14= 124Use case points (UCP) = 124 x TCF x ECF= 124 x 1.01 x 0.875= 109.383 or 109 Use Case Points

Duration = Use Case Points * ( 20 - 28) hours per week = 2180 to 3052 hours35

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Estimate Cost and ScheduleExplaining 20 28 hours estimate: According to approach proposed by Schneider and Winters (1998), they suggest counting number of environmental factors in E1 through E6 that are 3, and those in E7 and E8 that are below 3. Hence if the total is 2 or less, assume 20 hours per use case point, and for total 3 or 4 assume 28 hours per use case point. Any total larger than 4 indicates that there are two many environmental factors stacked against the project.[4, Page 8]Since our environmental factors total falls in two or less category, we assume 20 hours per use case point.Therefore, 109 use cases * 20 hours per use case point: 2180 hours (person hours)36

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Estimate Cost and ScheduleAccurate cost estimates are critical to both developers and customers. They can be used to request for proposal, contract negotiation and scheduling.We got 2180 hours (For 1 person). We estimate having 3 persons working per iteration and assume each iteration lasts two weeks: We also estimate that each developer will spend about 30 hours per week on project tasks. The rest of their time will be sucked up by corporate overheadanswering email, attending meetings, and so on[5]37

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

ScheduleMilestones of a project are the visibility of a project

We estimate 2180 person-hours of project work Therefore, for 3 persons working each week in a 30 hour week. 30*3 = 90 hours per week work., which makes an iteration(2 week) 180 hours2180/180 = 12.11 = 13 two week iterations= 26 weeks together

Therefore, 26 calendar weeks38

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Estimate Cost and ScheduleFor 2180 hours, 2180 * $100 = $218000, also we need to consider 30% overhead cost30/100 * 218000 = $65400218000 + 654000 = $283400Therefore, total cost of the project is : $28340039

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Schedule

PhaseStartEndAssociated MilestonesInception PhaseWeek 1Week 6Lifecycle objective milestoneElaboration PhaseWeek 7Week 15Lifecycle architecture milestoneConstruction PhaseWeek 16Week 22Initial operational capability milestoneTransition PhaseWeek 23Week 26Product release milestone

40

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Financial BenefitAccurate weather data is demanded by significant amount of users, for nearly every facet of our population, the weather data has critical role in decision making.Major buyers:Aviation forecastersMarine IndustryDaily commuting users 41

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Financial BenefitSome studies have examined value of short term weather predictions:Saving to oil drilling companies in Gulf of Mexico from avoiding unnecessary drill rig evacuations[1]Improved fueling decisions at Australian airports resulting from better forecast, could save companies $6 7 million per year[1]Better hurricane forecast for Atlantic coast over past 100 years have resulted in major reductions in yearly deaths in hurricane activity[1]42

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

SummaryAll sectors whether private or public get affected to sudden weather changes, having at least near to accurate prediction of weather can generate huge impact and reduce losses.With the help of this software we can work towards achieving this goalWe have prepared a business case which plans the cost estimated using the use cases laid out on basis of functional requirements and also estimate the schedule of the project. We also have designed a high level architecture and plan to proceed with vision artifact after getting approval for the business case43

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

References[1] https://www.gwu.edu/~spi/assets/docs/Socio-EconomicBenefitsFinalREPORT2.pdf[2]For cost estimation using use case: http://www.cs.cmu.edu/~jhm/Readings/Cohn%20-%20Estimating%20with%20Use%20Case%20Points_v2%2012-24-50-761.pdf[3]http://sceweb.sce.uhcl.edu/helm/rationalunifiedprocess/examples/csports/ex_sdp.htm[4]http://www.srh.noaa.gov/srh/dad/coop/USEWX.pdf[5] http://www.methodsandtools.com/archive/archive.php?id=25p344

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

ReferencesImage References:

http://www.slideshare.net/NitHik1/data-mining2-43891941http://www.zentut.com/data-mining/data-mining-architecture/45

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Weather Forecasting using Data MiningVision04/06/2016SWEN 5230-03Software Project Management46

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

OverviewStakeholder and User DescriptionsProduct OverviewProduct FeaturesConstraints47

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Stakeholder & User DescriptionsMarket demographicsDaily changing weather plays important part in everyones life. Primary users for the system are general public, aviation, fire and marineAviation forecasters use them in order to keep an eye on surface observations for wind shear, restrictions to visibility that could affect takeoffs and landingsMany users have been relying over different website for reliable weather service, and most of them get data from the pro National Weather Service, branch of National Oceanic and Atmospheric AdministrationEvery member of population uses weather data on regular basis, thinking of how weather can affect your travel, activity and business decision, the list of users become longer

48

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Stakeholder & User DescriptionsMarket demographicsNWS has been providing weather reports on county by county basis which on its site which is free, ad free and quick loadingThe other sites which followed the popularity were Weather Underground, Forecast.io, Weather Spark, Intellicast and not to forget the Google, as its easily available, but its forecast however is provided by Weather.com.[1]The users are anticipated to be consumers who already use cell phones on a regular basis for personal and/or business useAccording to report[4], the weather apps are the most downloaded app after gaming apps, they have generated the highest total revenue from paid downloads. Its really a must-have app for every smart phone user we want to know whether were about to walk into a tornado or not

49

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Market Demographics: Tim Spangler, director of the COMET atmospheric sciences training program, estimates the broad U.S. weather and climate industry at more than $5 billion, including some 250 commercial weather companies that generate roughly $2 billion [5]50

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Stakeholder & User SummaryNameRepresentsRoleKey NeedsPrivate Weather ServicesWeather data providersA single weather service cannot succeed on basis of its collected data, but requires contribution from different servicesThe system should accept this historical weather data which shall be helpful for developing patterns State and Local government agenciesWeather data providersAs we plan weather prediction on county basis, the local agencies should also be involved in data contributionThe system should accept local as well as global weather data which will be helpful in better prediction

51

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Stakeholder & User SummaryNameDescriptionKey NeedsWeather ResearchersEnd users of the systemThese stakeholders would be studying the predictions provided by our system hence we also can refer them as a stakeholderRequirement EngineersThis stakeholder works with customers and other stakeholder to translate needs into requirementsSpecifies domain, categories requirements into functional and non functional. Refines requirements as neededSoftware ArchitectThis stakeholder is the lead in development of prediction systemThe Architect will be required to be responsible for architecture of system, guides design and implementationProject ManagersThis stakeholder is the lead in development of prediction systemThe manager will be required to plan, manage, coordinate interactions and keep the team focused

52

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Stakeholder & User SummaryNameRepresentsRoleKey NeedsPublic SocietyEnd users of the systemDaily commuters, who rely on weather updates before travel plansThese will be the normal end users of the system relying for weather and storm updatesWeather broadcastersEnd user of the systemThese can be different web applicationAccurate Weather forecasts

53

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Product Overview

PerspectiveThe diagram explains an high level perspective of the product we plan to develop. The end user will input the current temperature and humidity into the system using mobile/desktop device. The system will have data set feed from the database which was entered by the Admin. The system then runs prediction algorithm in order to forecast the temperature and wind also will notify the end user with these values.54

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Product Overview The vision document applies to Weather forecast using data mining application which will be developed as a web application. The application will provide way to predict weather based on old weather data feed into system. User will enter temperature, wind and humidity and system will predict weather from previous data in database. Since the system will predicting based on the real old data fed earlier, the prediction can be reliable.

55

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Product OverviewSummary of capabilities

Customer BenefitSupporting FeaturesNear to accurate weather prediction daily commutersVoice based alerts for mobile appsWeather forecasts in advance for particular fly zones will benefit airplane companies to reroute planesContinuous alerts to Air Traffic Controllers Being able to pinpoint a wintertime low temperature can be the deciding factor for farmers in making decision regarding deployment or non-deployment of freeze prevention methods to save their cropsText message alerts to farmers in case of emergency weather situations

56

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Product OverviewAssumption and Dependencies:In using the onscreen keyboard, it is assumed that the user is literate and can typeThe default language for the application shall be US English. It is assumed that users who cannot speak and write in English will not be using the text to speech features(third party mobile app) in the system, at least initially.The prediction would be dependent on the old weather data provided, and may not give accurate results every timeThe device used by end user mobile or desktop is required to be connected to the internet in order to access the weather data

57

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Product FeaturesFeatureUser or Stakeholder Need MetAbility to allow administrator to feed historical data This meets the stakeholders need to make use of the historical weather dataAbility to hold user contact information in order to send alertsThis meets the users need to be notified about weather prediction on provided contact informationAbility to recognize pattern in temperature, humidity and wind from old dataThis meets the stakeholders needs to apply algorithms and recognize patternsAbility to apply prediction algorithms and store results to databaseThis meets stakeholders need of preserving data

58

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Product FeaturesFeatureUser or Stakeholder need metAbility to run reports and plot graphs of weather conditions predictedThis meets the users requirements of having the prediction in form of visuals on map and Ability to export data in order for third party systems to use itAbility to notify registered users about daily alertsThis meets the users need to be notified about weather prediction on timely basis

59

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

ConstraintsUsability:Ability to read maps, as the weather pattern would be plotted over mapsMust be able to understand units of temperature(F or C), wind and humidityPerformanceAccuracy of result depends on accuracy of previous weather data fed into systemAlgorithm calculation may take time, hence delay for data to load

60

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

SummaryOur vision of project is develop near to accurate weather prediction system based on historical data.We listed the number of stakeholders and users who are the direct beneficiary of the weather prediction results.Our every listed feature focuses on satisfying the present need of weather forecast, as we march towards our goal we seem to suffer certain constraints which we hope to overcome.61

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

References[1] http://lifehacker.com/5897973/five-best-weather-web-sites[2] http://www.trishmarie.com/vision.pdf[3] http://www.livingontherealworld.org/?p=1317[4] http://techcrunch.com/2011/09/12/report-android-market-nearing-6-billion-downloads-weather-apps-are-makin-it-rain/[5] http://www.cnbc.com/id/4367283962

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Weather Forecasting using Data Mining

ArchitectureApril 12 2016SWEN 5230-03Software Project Management63

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

OverviewArchitecture goalsUse case viewLogical viewProcess viewDeployment viewImplementation view64

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Architecture GoalsThe system should allow administrator to add historical weather dataThe system should be able to recognize patterns in temperature, humidity and wind with use of historical dataSystem should periodically apply prediction algorithms or models on obtained data and store results to central databaseSystem shall obtain and display confidence value for each prediction given to user

65

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Architecture GoalsSystem shall support registering new user by providing email, so system can send daily weather updates to registered usersSystem shall allow users to check weather for one week aheadSystem should be able to alert its registered users by sending storm warnings in emergenciesSystem should have 99.99% uptime and should recover and be running immediately after disaster An uptime guarantee would be a good indication of our applications commitment to keeping our application up and running

66

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

High Level Use Case View

67

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

High Level Use CaseThis is an high level use case diagram which explains the overall high level functional requirements in our systemThe user, admin and database are the primary actors interacting with the systemThe user first registers himself/herself in the system with email id for notifications, and may also enroll its phone number for instant alertsThe Admin is responsible for continuously feeding historical weather data into the system from the databaseThe system holds prediction algorithms which shall study patterns in the data and provide future predictions of weather to the user

68

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Use Case View Machine side

69

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Use case view Machine side The machine side use case view provides overview of major functional requirements on the machine side developmentMost of the use cases are the stages involved in data miningThe Admin feeds the historical weather data into the system, then under the data preparation, we conduct data staging, where data is prepared and kept ready to mine. The data staging area sits between data source and data targets. Further, data is cleaned and redundant data is removed.The cleaned data is now processed by data prediction algorithmsAlgorithms study the data for patterns developed in data sets, and the ready weather prediction is provided to the end user

70

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Logical View(Class Diagram)

71

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Logical View (Class diagram)The Logical View class diagram provides the necessary classes working in our system and their relationships.An User can be a registered user or a visitor to our website, hence their respective classes handle their personal information attributes.An registered user holds an user account, hence user account is a class which handles registered user account informationA user checks the weather information hence, weather information is also a class which will hold the weather data to be shown to the end userThe weather information would be pulled from the database which also is a class

72

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

High Level Process View

[1]73

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

High Level Process ViewThe High Level process view diagram is flow of process in our system as the diagram explainsThe weather data is collected from different weather stations and collected at the data centersThe data experts at the data centers analyze the data and feed into the data prediction modelThe prediction reports are generated and the forecast is shown using graphical representation on maps74

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

End to End Process View

[3]75

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

End to End Process ViewThe End to End Process View diagram explains the process in our system in more detailed manner.Starting with the data gathering, the data is prepared and filtered furthered The cleaned data is stored in database and using prediction algorithms the data pattern is analyzed and reports of the forecast are generated

76

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Deployment View

[2]77

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Deployment ViewThe deployment diagram shows the hardware for our system, the software that is installed on that hardware, and the middleware used to connect the disparate machines to one anotherThe Presentation Server is the user end, it can be the device which has a browser to display the resultsThe Presentation Server communicates with the Data Mining Application Server which has components like the warehouse for storing the ready data, the data miner and the pattern evaluator component who works on the data.The Data Mining Application Server communicates with Database Server to grab the historical weather data in order to operate over it.78

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

SummaryA well designed architecture is a major milestone in success of software project and hence we make an effort to architect the design view, logical view, process view and deployment view.The design view have completed using use case diagram, which describes an high level and a machine perspective use case.The deployment view at the end describes the hardware and software layout expected when product is ready. 79

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Image Referenceshttp://www.ctwatch.org/quarterly/print.php%3Fp=103.htmlhttp://agilemodeling.com/artifacts/deploymentDiagram.htmhttp://www.slideshare.net/jafaeldon/data-centric-hpc-for-numerical-weather-forecasting80

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Weather Forecasting Using Data MiningWork Breakdown StructureApril 20, 2016SWEN 5230-03Software Project Management81

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

WBSAccording to paper Planning successful data mining projects by IBM, there are four critical data mining success factors for successful planning for data mining projects the right application, right people, right data and right tools, accordingly we have build this Work breakdown structure. [1]82

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

WBSThe above diagram shows overall layout of Work Breakdown Structure for our project Weather Forecasting using Data miningAll the phases the project must pass through for successful execution have been included, and every entity further breaks into manageable work packages.

83

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

WBS (Requirement Gathering)Our first major deliverable of the project is Requirements Gathering, the major component deliverables under this are gathering requirements from stakeholders and documenting them using use case scenarios and planning the activity flow using diagrams.

84

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

WBS(Planning)The project is a data mining project and hence our major focus here will also be on collecting data sets for carrying mining over it.The next major deliverable here is Planning, once the requirements for building the project have been gathered, we now plan to create a plan with strategic end in mind, for which we have to plan use of essential resources, the right data, right people, clear vision and plan data resources for future as well.Estimating project schedule and project milestones will be essential for monitoring the progress of our project.85

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

WBS(Planning)

86

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

WBS(Planning)Responsibility of the data expert would be to discover what relevant data is available to meet our strategic business goal and to help us identify the right data by answering questions such as: What do the codes in this field mean? and Can there be more than one record per weather parameter in this table? Consider how much data knowledge is available and evaluate any risks caused by its absence or scarcity.Under building the right team deliverable, we primarily focus on hiring or training a Data Miner and Data Expert as our project revolves more around working on right data. A Data miner is the one who understands the questions that can be answered using data mining and how to answer them while a Data Expert is the one who understands existing IT capabilities and signs off on data mining integration with those systems[1]

87

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

WBS(Design)

88

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

WBS(Design)Next major deliverable is Design:This deliverable holds most important component deliverables which provides a blueprint of the system which will be developedArchitecture will hold information regarding the correct software development lifecycle picked and the flow of activities expected within its various phasesThis involves creating structural models like class diagrams which covers the major classes and its associations in system and can be beneficial for developers who can directly convert them to actual classes in program also the deliverable component includes behavioral models like activity diagrams which explains flow of information among various components in the systemThe design deliverable also includes creating a prototype of the expected software and getting it approved in stakeholders meeting so that desired changes can be implemented immediately in prototype and same be documented for product development89

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

WBS(Execution)

90

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

WBS(Execution)The next major deliverable is Execution where the we actual development is planned to happen The requirements collected until now are verified and validated and the development is initiated, development is initiated in two stages:Client side development:Build interface for users to check weather information by input of current temperaturesBuild maps which show weather resultsBuild graphs which provide easy analysis of future weather predictionsServer side developmentBuild interface for administrator to enter historical weather data setsRun prediction algorithmsContinuously update system with new weather data from different weather sources via APIs91

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

WBS(Testing)

92

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

WBS(Testing)The testing phase starts unit testing of code packages followed by testing of whole after units are integrated.The environment testing is carried out where system is tested for entered data against production data and in different environmentsThe next testing deliverable includes work packages for assessing the developed product under various user scenarios The system undergoes testing with selected set of users, they can be registered and non registered usersThe use of system is documented called user manual for end users, also required training is provided internal to company and the product then goes on production

93

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

WBS(Control)

94

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

WBS(Close-out)

95

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

WBS(Control and Close-out)The control deliverable is the project monitoring entity which will hold work packages for status meetings and continuous update to project plans as the plans keep changing and the project keeps progressingEvery iteration release is monitored by Project managerAfter approval is received from tester the releases at development/integrations stage are promoted to production by project managerThe close-out is the summarization of the project, which includes documenting the software, gain formal acceptance of software created from the sake holders

96

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

WBS(Control and Close-out)When all the releases over the production are approved by the tester over production, the manager promotes and the system is deployed to the users end97

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Summary98We have designed the work breakdown structure, and laid out major components of our project at level 1 and further include the small work packages under every component at different levels.

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

References[1]. Planning successful data mining projects http://www.besmart.company/wp-content/uploads/2014/12/Planning.pdf99

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Weather Forecasting Using Data MiningSoftware Development PlanApril 27, 2016SWEN 5230-03Software Project Management100

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

OverviewScopeOverview & objectivesAssumptions and ConstraintsDeliverablesProject OrganizationStaffingControl & PerformanceMilestones101

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

ScopeIn this project we plan to investigate use of data mining in forecasting temperature, rainfall, evaporation and wind speed using historical weather data This Software Development Plan describes the overall plan to be used in our project, including deliverables of the product. The details of the staffing will be described in the staffing table. The plans as outlined in this document are based upon the product requirements as defined in the Vision Document[3].

102

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Overview and ObjectivesThis project will implement a customized for Weather forecast. This will deliver notification of weather report to registered users cellular phones or electronic mail while unregistered users can check the weather updates on the landing page of the website. The users will be required to enter todays weather information in order to receive future forecast using data mining.103

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Assumptions and ConstraintsStaffing: We do not have a release/delivery manager and those duties are preformed by project manager, which is a staffing constraint. Additional open source contributors are welcome, but are not assumed.Budget: Personally financed Schedule: Development efforts are constrained by "day-jobs". 104

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

DeliverablesBusiness Case: The business case will be delivered by the project manager to the stakeholders of the project during the week 01 Inception phaseVision Artifact: The vision artifact is delivered by project manager to stakeholders during week 02-03Requirements Gathering: The requirements gathered with the use cases and user stories will be delivered to the project manager from requirements specialist during week 03 04 Software Development Plan: This is the final draft of the development plan before the actual development starts and is delivered by project manager to stakeholders during week 05 -06.

105

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

DeliverablesSupplementary Specifications: The Supplementary Specifications shall capture the system requirements that are not readily captured in the use cases of the use-case model and which shall include legal and regulatory requirements and should be delivered by the business analyst and requirements specialist to developers during week 03 04. Developers will use the Supplementary Specifications as a reference when defining responsibilities, operations, and attributes on classes, and when adjusting classes to the implementation environment[1]Creative Design Briefs: The designer will deliver the design briefs to stakeholders and after getting approved from them the developers should start development over same under consent of project manager during week 07 - 08.

106

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

DeliverablesUser Interface Prototype: The user interface prototype is designed by developer and delivered to stakeholders during week 09- 10 for approval meeting before initiating with development in projectHistorical Weather Data Gathered: The system uses data sets of weather report of past few years which are delivered by local and state government agencies monitoring weather during week 05-07 and build patterns of changing weather and with help of algorithms predict the future weather prediction. Data Model Design: Since the systems efficiency primarily revolves around the use of data, it is very necessary to have a clear and efficient data model design. The data model design will be delivered by the database developer to project manager, and get it approved before week 15 107

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

DeliverablesModel Database : After the design for database is ready, the database model is delivered by database developer to himself/herself during week 15 and the historical weather data is fed into database by database developer.Design Software Architecture Document: The software architecture document is delivered by the software architect to project manager during week 12. Under which, the right SDLC lifecycle is chosen which fits the system planned to build.108

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

DeliverablesTest Package : These are the test cases designed by the testing developers to themselves during end of week 23 and delivered in order to test thoroughly the developed system.Change Requests: The system after undergoing through testing might need changes which are now delivered by testers to developers post week 23. Test Summary: The system might undergo numerous changes recommended by testers and after the changes are delivered, documentation of the completed changes are made which is delivered by testers to project manager end of testing phase and delivered during end of week 24109

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

DeliverablesImplementation Subsystem: The development of the system starts and is delivered by the front end developers and database developers to project manager and end of week 21. User Documentation: As the project progress and the system evolves, the features are delivered by the project manager to the ready to ship software package with inputs from the team and the complete is user documentation is delivered at end of the project that can be week 25 - 26.

110

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Project OrganizationWe plan to follow more of Project based organization structure than a Functional based organization structure as traditionally followed in larger organizations.

We plan to have integration and development environment as same with use of SVN(Subversion)111

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Project OrganizationIn the project organization shown graphically in the earlier slide, has project manager as the leader who leads the project and monitors the other teammates performance, schedule and the team directly reports to the managerThe Data Expert and Data Miner are exclusively needed for this project as we need expert to gather and study data before feeding to our system. Further, the data miner is responsible for studying patterns in data. They both will be part of the database team, but not necessarily reporting to database developer

112

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

StaffingRole ResponsibilitySkills RequiredNumber of staff requiredDurationProject ManagerLead Team, Report statusProject Management1Complete durationRequirements SpecialistsRequirements gatheringResearch and Analysis1Week 1 Week 7Business AnalystRequirements validations Presentation, Communication and Critical Analysis1Complete duration

Web DeveloperCreate framework contentWeb developer1Week 30 Week 50

InceptionElaboration113

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

StaffingRolesResponsibilitiesSkills RequiredNo. Of Staff RequiredDurationData ExpertUnderstand existing IT capabilities and sign off on data mining integration with systemsData Extracting and Data Loading1Week 30 Week 50Data MinerBuild algorithm to understand pattern of data fed by Data ExpertData Analysis1Week 30 Week 50

114

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

StaffingRolesResponsibilitiesSkills RequiredNo. Of StaffDurationTesterThorough testing of software Building detailed test cases and testing software in different scenarios1-2Week 51 Week 77Software ArchitectDevelop suitable SDLC model for the systemArchitecture Experience1Week 8 Week 29

115

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Control and PerformanceSchedule & Budget control plan: Project status reports will be issued weekly and will include Milestone Tracking detail to ensure that the project stays on track. Changes in the schedule will be escalated to the project manager, who will then decide whether to alter scope in order to preserve target completion dates.[1]Quality Control Plan: Formal reviews will be executed for each design and implementation subsystem post every iteration in presence of manager. This will ensure that the objects under review meet the specified requirements.[1]Close-out plan: At the end of the project, a Lessons Learned meeting will be held to capture new techniques, tools, or methods. Project deliverables will be archived to the Knowledge Management repository for future reference.[1]116

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Control and PerformanceMonitoring Performance: We have laid out certain Key Performance Indicators(KPIs) which will help us in measuring how the project is performing. KPIs are not objectives but the readings that enable a manager to assess performance towards the achievement of objectives.[2]Typical KPIs we use will be schedule and budget compliance, number of scope changes, number of issues and defects, and stakeholder satisfaction.Projects must end and completion time is regularly firmly connected to the business targets that drove the project's introduction. Following to a financial plan in dollars and/or asset time is a key pointer since it gives us a feeling of whether we are executing as we have anticipated that would perform. Much of the time, project financers and customers are cost conscious. They need to know the amount they will spend on the venture and they need to know it before the undertaking is over.117

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

Milestones

PhaseStart DateStartEndEnd DateAssociated MilestonesInception PhaseJuly 01 2016Week 1Week 6August 12 2016Lifecycle objective milestoneElaboration PhaseAugust 15 2016Week 7Week 15October 10 2016Lifecycle architecture milestone

118

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

MilestonesPhaseStart DateStart WeekEnd WeekEnd DateAssociated MilestoneConstruction PhaseOctober 11 2016Week 16Week 22December 06 2016Initial operational capability milestoneTransition PhaseDecember 07 2016Week 23Week 26December 28 2016Product release milestone

119

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

References[1]http://sce.uhcl.edu/helm/RationalUnifiedProcess/process/artifact/ar_sspec.htm[2]http://www.projecttimes.com/georgepitagorsky/measuring-in-progress-project-performance.html[3]https://kenai.com/projects/cps/pages/SoftwareDevelopmentPlan#Scope120

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar

SummaryWe have worked with all software artifacts, and did encounter a lack of abstraction; but have attempted to deal with all details of the artifacts. Examples assuming all small work packages that would be needed to complete and adding them in WBSIn this project, we are investigating approaches to generate reasonable summaries of software artifacts. The goal is to raise the level of abstraction and improve the productivity for software developers.With an estimated plans earlier discussed for schedule, staffing and cost estimation using use cases, we have sufficient information to proceed with our Weather Forecasting Using Data Mining project.

121

SWEN 5230-03 Software Project ManagementRushikesh Mangrulkar