==Attendance==
Congyang Peng
Kristine Gloria
Rui Yan
Matt Ferritto
Christopher
Randy
Xiaochuan Ma
Liying Zhang
Qi Pan
Nupoor Deobhakta
Ruiwen Liu
==Notes==
Hands-on session with Amar:
If you have specific questions, email him () and CC Prof. McGuinness, Patrice, and Katie, please~
Monday (after 4) -
Tuesday -
Wednesday (10am in the Winslow lounge) - Kristine, Ray, Chris, Matt
Thursday or Friday - Congyang
**Can the First Responder group volunteer to go first for presentations? - Kristine
Gourp one -- CarFire
Created hierarchical structure base on Fire service hand tool ppt by google search.
Classifier were created by reasoning? or just given? Given
Enable reasoning like: if a car belongs to Electricity car but not a Hybird, it won’t comsume gasoline...
FireExtingulsher is an instance of ClassCFireExtingulsher
In future could create more Extingulsher
It’s OK it the group don’t give beautiful visulization. But the group should highlight the usage and importance of semantic tech.
Show more semantic value of the project, the reasonser should be able to recognize (which is to infer) that electric fire is a “ClassCFire”
The rabbit tool and hydraulic door opener is the same. but still need to let the reason infer that rabbit tool is the same as hydraulic door opener and notice that rabbit tool is a jargon.
-- use “sameAs” for class to class
Group two -- Hospita.ly
Use bar chart represent the percentage rate that if the hospital have a good nurse communication rates.
Hospital: 4 major classes: Best/Good/ Average/ Bad …
Best means mortality and remission rate are both than average.
Allow user to add more constraints such as to nurse communication rates.
About the ontology, how about creating perspectives such as: mortality perspective/ remission rate perspective...
Enable predefined concept to be modified. For example, from my point the distance is more important than remission rate.
Follow-up : Add more sub-category in the hospital ontology
Also, pre-define good hospitals for the users from a broader range of tastes
Consider more uses of ICD 9 and SNOWMED existing ontology
Group three -- SemantEco
Water quality data - converted to RDF, all from USGS; need to identify what data are new and what are duplicates.
USGSP-code gives characteristic (eg, Boron) and units (eg, ug/L) - is there a way to figure out if two measurements are the same (taken in the same place at the same time) but in different units?
Water ontology is already in place, so nothing new now. Right now, mostly adding new water data → focus on what can be done to add more value: maybe add more with the provenance/consistency checking; or, look at more contaminants; or, look at the unused fields given in the data, since many seem to have been ignored (collection/testing method? ground water vs surface water vs sediments?). If not within the scope of this project, discuss what could be done in the future.
Air data needs to build new ontology parts.
Air quality data - converting to RDF, from EPA; need to define equivalent properties so they are properly recognized by the conversion tool.
also incorporating site IDs into the ontology; also extending the SemantAqua; adding Excessive Measurements, and AQI categories. Working on converting the rest of the air data, and combining it with the site data.
Different rules for air contaminant measurements have different units - currently they are specified in the rules themselves
For AQI, scope and look (at least) specifically at O3 or PM10 - cause problems with asthma
- demonstrate inference: if CO has a Hazardous AQI, it can also be inferred that it is violating the EPA standard
- some measurements are over an 8-hour period (avg. level must exceed a threshold over an 8-hour period); some regulations have 1-hour and 8-hour levels?
Bird data - using Geospecies Ontology vocabulary, and making appropriate enhancements/conversions. → This makes the search easier by imposing a hierarchy; makes implicit information explicit.
currently working on converting data: finished Great Backyard Bird Count (from USGS); just got data from eBird, need to convert FeederWatch data, but it can be done in a similar way.
UI design - jstree for tree structure nodes, very similar to ontology
Can currently search the tree for bird species names to see the data.
Working on figuring out how to integrate this information into the website (not plotting locations right now on the map?) - scope it so you can plot something
Working on client-server communication so that user can select a zip code and see species information available OR if user searches for a specific species within a zip code and no results or returned, it can suggest a related species.
Integrated google maps API v.3
Created icons for air data to differentiate from water data; allow users to click and see data and visualize data (still need to implement a threshold line on visualization to mirror SemantAqua). Also working on displaying species data alongside the air data for each location)
Right now, still using fake/placeholder data - close to bringing in actual data, then the group can work on constraining the data that is displayed. Also working on unifying air information with existing water information.
Lecture:
RDFS -- ontology language
weak in describing details
weak in reasoning
OWL -- identified for Web Ontology Language
possible for providing reasoning support
more expressivity compares to original version
OWL: only one fragent (OWL Lite)
OWL2: better datatype support
three fragments with computational properties
SKOS
RIF
SPARQL
OWL-S
Provenance -- where to get the data, and you encoding that information
knowledge provenance; enrich with ontologies
JUST AN ASIDE: