Transcript Time Series Analyst An Internet Based Application for

Time Series Analyst
An Internet Based Application for Viewing
and Analyzing Environmental Time Series
Jeffery S. Horsburgh
Utah State University
David K. Stevens
Utah State University
Jon Goodall
Duke University
The Problem
• What is the spatial and temporal distribution of
data available for scientific or management
studies?
• How do we assemble and explore environmental
time series data?
– Many different sampling programs, agencies, etc.
– Many different sampling locations, frequencies, etc.
What is the Time Series Analyst?
• Provide a window to explore
the available data
– Exploratory Data Analysis
• Distribution (spatial and temporal)
• Density
– Plotting Data
– Generating statistical summaries
– Simple means to “slice and dice”
the data
Time Series Analyst
Plot Types
Plot Options
Plot Window
Summary
Statistics
Station and
Variable
Selection
Date Range
Selection
How have we used Time Series Analyst?
• Watershed water quality studies and
TMDLs
– Management of water quality data
– Generation of data summary reports
– Delivery of water quality database AND
visualization tools
Time Series Analyst Features
Time Series
Probability
Histogram
Box and Whisker
Monthly, Seasonal, Annual,
and Overall
Original Time Series Analyst
• Simple, map based point and click access to data
Original Time Series Analyst
• MapWindow Plug-in
• Development
Environment
– Visual Basic .Net
• Plotting Control –
Gigasoft
ProEssentials
http://www.gigasoft.com
• Time series data
stored in Microsoft
Access or SQL Server
relational database
MapWindow
Personal Computer
Access or
SQL Server
Database
ProEssentials
Plotting
Control
Time Series
Analyst
Issues and Limitations
• Requires Software Installation
– Database updates
– Software updates
• No facility for realtime or continuous data
because database is essentially static
First Internet Based Version
http://water.usu.edu/analyst/
Web Browser
Client - anywhere
internet connection
is available
Local SQL
Server
Database
Time series data stored
Locally on USU server
Internet
Web Server at USU Running
the Time Series Analyst
First Internet Based Version
• Development Environment
– Microsoft ASP.Net
– Microsoft SQL Server
• Added capability to incorporate realtime
sensor data
• Addresses issues with client software
upgrades
• Coupled with ArcIMS map server to
preserve map linkages
SQL Queries
passed from
Time Series
Analyst to the
server database
User Interaction
through Web Browser
Time Series Data
Stored in Microsoft
SQL Server Database
Query results are passed back to the Time Series
Analyst where they are plotted and displayed in the
browser
Query results can be
exported to a browser
window or directly to
Microsoft Excel
How Do We Store and Serve
Disparate Monitoring Data?
• Robust
• Interactive
• Simple…
• Core Tables
– Stations
– Parameters
– Data
Original Relational Database
Storing Disparate Monitoring Data
HODM - A More Robust Schema
• One database schema
to store all
observational data
• CUAHSI Hydrologic
Observations Data
Model (HODM)
–
–
–
–
Generic schema
Stores metadata
Data versioning
Provenance of data
Using a Served Database Approach
• Advantages
– All types of data under one roof
– Dynamic – can be inserting data at the same time it is
being queried out
– Simplifies data access queries
• Disadvantages
– Design - Will one schema really store all of the data?
– Implementation - Not all DBMS’ are free
– Management - Burden to ensure most recent data
CUAHSI NWIS Web Services
http://river.sdsc.edu/NWISTS/NWIS.asmx
• Machine to machine
communication of
data over the internet
• Users can program
against NWIS as if it
were on their local
machine
• Replace SQL queries
to database with calls
to the appropriate web
service
Web Browser
Client - anywhere
internet connection
is available
USGS server with national
NWIS data repository
Internet
Internet
Server At USU running
the Time Series Analyst
Internet
Server at SDSC running
the CUAHSI NWIS Web Services
Web Services Based Time Series Analyst
http://water.usu.edu/nwisanalyst/
• Advantages
– No database for us to maintain!
• Doesn’t preclude having a local database…
– Provides access to any USGS site in the
NWIS repository!
• Disadvantages of Web Services
– Speed
– Limited Query Ability
Parameters
• Parameters that can be
passed:
–Station Name
–Variable/Parameter
–Start Date/End Date
–PlotGraph = True/False
http://water.usu.edu/nwisanalyst/default.aspx?Database=
WQ&Station=10109000&Variable=00010&StartDate=01/0
1/1975&EndDate=12/31/1994&Plotgraph=True
Conclusions
• Stand alone applications and databases can be
useful, but they are static
• Server based software and databases (HODM) may
be the answer for our own data, but people may prefer to
get data direct from national repositories rather than our
copy
• In terms of Hydrologic Information Systems:
– A combination of server based instances of HODM and web
services for accessing national datasets may be the best way to
go
– Applications like the Time Series Analyst are needed to provide
users with the ability to “wade through the data”
Acknowledgements
• David Stevens – The Father of Time
Series Analyst
• EPA Targeted Watersheds Grant – Bear
River Basin
• CUAHSI HIS Project
• EMRG Programming Team