MSR Internship 07 Whats cooking Jayant Gupchup MSR

  • Slides: 11
Download presentation
MSR Internship 07 What’s cooking ? Jayant Gupchup MSR Intern 07, Ph. D. ,

MSR Internship 07 What’s cooking ? Jayant Gupchup MSR Intern 07, Ph. D. , The Johns Hopkins University BARC, 07

Outline • Part 1: Ameri. Flux – Help Ameri. Flux scientists browse “the cube”

Outline • Part 1: Ameri. Flux – Help Ameri. Flux scientists browse “the cube” (No more MDX, really ? ? ) – Analyze the data in Matlab (Let’s face it. . Matlab is more powerful than excel) • Part 2: Russian River – Apply Spectral and Pattern recognition techniques to try and answer Rebecca’s questions. BARC, 07 2

Ameri. Flux MDX ? ? MDX Builder Query Cube (Get the data out of

Ameri. Flux MDX ? ? MDX Builder Query Cube (Get the data out of the cube) Render Cube Results Plot Results using “Matlab” BARC, 07 3

MDX Builder What you want ? Group by About whom you want ? What

MDX Builder What you want ? Group by About whom you want ? What period ? MDX : SELECT NON EMPTY CROSSJOIN ({[Datumtype]. &[32], [Datumtype]. &[8]}, {[Site]. [Veg. Type]}, {[Site]. [Longitude]}) DIMENSION PROPERTIES PARENT_UNIQUE_NAME ON COLUMNS, NON EMPTY CROSSJOIN ({[Timeline]. [Year]. &[2000]: [Timeline]. [Year]. &[2006]}, {[Timeline]. [Day of Year]}) DIMENSION PROPERTIES PARENT_UNIQUE_NAME ON ROWS FROM Bev. Law. Cube WHERE {[Measures]. [GPPOver. LE]} BARC, 07 4

Dimensions (A Pain in the ‘@$$’) MDX BARC, 07 5

Dimensions (A Pain in the ‘@$$’) MDX BARC, 07 5

Query the cube & Render MATLAB ASP/C# MDX Query HTML/ASP Query Results “Serializer” Matlab

Query the cube & Render MATLAB ASP/C# MDX Query HTML/ASP Query Results “Serializer” Matlab reads HTML “Deserializer” MATLAB OBJECT : rows: 1096 cols: 10 dim 1 axes: 3 dim 2 axes: 2 dim: [1 x 2 struct] data: [1096 x 10 double] axis: [1 x 3 struct] BARC, 07 axis: [1 x 2 struct] 6

Plots using Matlab BARC, 07 7

Plots using Matlab BARC, 07 7

To Do (Connect the dots !) Cube Reader MDX Builder Cube Renderer Matlab Plots

To Do (Connect the dots !) Cube Reader MDX Builder Cube Renderer Matlab Plots BARC, 07 8

Russian River (So far) • Build Flood DB: – Identify interesting sites & ones

Russian River (So far) • Build Flood DB: – Identify interesting sites & ones with good data – Compute Flood cut-off values for different sites – Associate the same flood at different sites • Time Delay • Overlap between floods. BARC, 07 9

The “Cool stuff” • Address some science questions: – Correlate Discharge, Embeddedness with •

The “Cool stuff” • Address some science questions: – Correlate Discharge, Embeddedness with • Turbidity • Fish density / distribution • Landuse – 1. 5 year flow analysis. – Identify patterns and trends in Flow statistics such as <E 1. . E 4>, Time 2 Peak, Area. Under. Flow etc. – Classify flood origins using discharge signatures. BARC, 07 10

Discussion BARC, 07 11

Discussion BARC, 07 11

Jayant Gupchup Graduate student Johns Hopkins University RepresentativeJayant Gupchup Graduate student Johns Hopkins University Representative
Statistical challenges In Sensor Networks Jayant Gupchup AdvisedStatistical challenges In Sensor Networks Jayant Gupchup Advised
Cooking Techniques Seasonings and Flavorings Cooking Cooking isCooking Techniques Seasonings and Flavorings Cooking Cooking is
Cooking of Food Cooking Cooking of food isCooking of Food Cooking Cooking of food is
Dry Heat Cooking Techniques What is Cooking CookingDry Heat Cooking Techniques What is Cooking Cooking
Cooking Methods Cooking Cooking can be defined asCooking Methods Cooking Cooking can be defined as
Melez Msr Islah Msr 2 n20 zerinde enMelez Msr Islah Msr 2 n20 zerinde en
MSR Effects Upon Secondary Cycle MSR Effects UponMSR Effects Upon Secondary Cycle MSR Effects Upon
Management Spirituality Religion MSR MSR Objective The overallManagement Spirituality Religion MSR MSR Objective The overall
Msr Zea mays Trkiyede Silajlk Msr Ekim AlanMsr Zea mays Trkiyede Silajlk Msr Ekim Alan