konza prairiekonza prairie long-term ecological research (lter)lter henry mikhail

27
PRESENTATION 3 Konza Prairie Long-Term Ecological Research ( LTER ) Henry Mikhail

Upload: candace-wilkinson

Post on 28-Dec-2015

215 views

Category:

Documents


0 download

TRANSCRIPT

PRESENTATION 3

Konza Prairie Long-Term Ecological Research (LTER)

Henry Mikhail

OUTLINEAction Items

Component DesignFormal SpecificationTesting EvaluationProject EvaluationFuture WorkDemoQuestions and/or comments

ACTION ITEMSOntology Visualization

A. Change the ontology layout; when a user clicks on a concept view related criteria in an new interactive window instead of a dropdown list.

B. Add functionality to create dynamically JSON file for the new interactive layout of the criteria.

COMPONENT DESIGN

Changed component names.Added one more operation to the ontology Component (an operationThat created dynamically a JSON file).

FORMAL SPECIFICATION-- $ProjectHeader: use 0.393 Wed, 16 May 2007 14:10:28 +0200 opti $

model Konza-- classesclass Get Data

operations getTableName(concept: String , criteria: String) : String

get Data( table Name: String) : Set(String)

createCSVFile( data: Set(String)) : Boolean

get Criteria(concept: String) : Set(String)end

FORMAL SPECS - CONTINUE constraints context GetData::getTableName(concept : String , criteria : String) : String pre validParams: concept > ' ' post validTable: criteria > ' '

context GetData::getData(tableName : String) : Set(String) pre : tableName > ' ' post : result-> size > 0

context GetData::createCSVFile(data: Set(String)) : Boolean pre : data->size > 0 post : result = true

context getCriteria (concept: String) : Set(String) pre : concept > ' ' post : result->size > 0

FORMAL SPECS - CONTINUE

context GetData::createJSONFile(concept: String, criteria : Set(String)) pre : concept > ' ' pre : criteria->size > 0 pre : criteria->forAll(c:String | c > ' ')

TESTING EVALUATIONUse cases

Interactivity and navigability

Data retrieval performance

User Interface performance

TESTING EVALUATION - CONTINUE

Use Cases

Tests were performed to verify that use cases were working as expected. I tried to test some of the concepts and their related Criteria. Unit testing was performed using assertion to verify the correctness of the java methods.

Interactivity and navigability

Actions that occur upon mouse clicks on a concept or criteria.Considering some of the limitations of this visual library. Possible action in every state of the navigation.

TESTING EVALUATION - CONTINUE

Data Retrieval Performance

Several data retrieval performance tests have been done to evaluate the response time of the application. It’s is clear that the amount of data, in terms of number of columns and rows retrieved. I will list the tests that have been done and analyze the results.

TESTING EVALUATION - CONTINUE

Test cases Number of rows: ~ 256 K Time: 5 seconds Number of rows: ~ 512 K Time: 8 secondsNumber of rows: ~ 767 K Time: 10 seconds Number of rows: ~ 1Million Time: 13 secondsNumber of rows: ~ 1.28 Million Time: 15 seconds Number of rows: ~ 1.5 Million Time: 20 secondsNumber of rows: ~ 2 Million Time: 23 seconds

INTERFACE PERFORMANCE

I performed several tests regarding this concept and I found out the maximum number of elements has to do with the location the distribution of the elements in the ontology branches. My final conclusion is that the maximum number should not be greater than 75 elements.

INTERFACE PERFORMANCE – CONT (65)

INTERFACE PERFORMANCE – CONT (75)

INTERFACE PERFORMANCE – CONT (75 L1)

INTERFACE PERFORMANCE – CONT (85)

INTERFACE PERFORMANCE – CONTTime required for the layout to stabilizeTests40 elements 4 seconds 55 elements 4 seconds65 elements 5 seconds 75 elements 7 seconds85 elements 8 seconds

TEST PLANUse cases

View ontology: Navigate to the url of the web page that shows the Konza ontology.

Search Data: Click on an ontology concept to test the search web page. Perform several test cases.

Download data: Attempt to download data files and save them locally.

PROJECT EVALUATION

Estimates Phase SLOC

Phase 1 100

Phase 2 650

Phase 3 250

Project Duration

Phase Estimated End Date

Actual End Date

Phase 1 01/20/2012 01/20/2012

Phase 2 03/16/2012 05/02/2012

Phase 3 05/01/2012 06/21/2012

PROJECT EVALUATION - CONT Time spent in each phase

PROJECT EVALUATION - CONT

Phase 1

PROJECT EVALUATION - CONT Phase 2

PROJECT EVALUATION - CONT Phase 3

PROJECT EVALUATION - CONT

Lessons Learned

Ontology creation requires lots of understanding of the domain.

Data visualization and JavaScript.

Dynamic creation of JSON files for data visualization.

Data import and format ( data cleaning).

FUTURE WORK

There are lots of aspects that can be expanded in the project;

the data retrieval process can be improved by having access to

real-time data instead of downloading data files and loading them

into a different database.

DEMO

QUESTIONS/COMMENTS