Landing the Raven: Positioning the Knowledge Discovery System in the Enterprise Wendi Pohs, Iris Associates
Contents What is the Knowledge Discovery System? Knowledge Management Architectures Content stores: Spiders Information warehouse: The K-map Classification: K-map Builder Retrieval: K-map Indexer Presentation: K-station Association: Metrics Tales from the trenches
– information, task aggregation – selection and display tools – people/place awareness – place creation and management Can work together or independently The Knowledge Discovery System Has Two Product Components – search and browse – taxonomy generation, concept clustering – expertise profiling and location – metrics
What does the K-station do? Place Management – Personal and Shared places May Include discussion forums, teamrooms, doc libraries, task lists, e- mail, MS Office integration – Manage People: Directory integration, security, membership, online awareness, realtime communication Integrates with the security and data model of Notes/Domino
A Customized K-station Place
K-station - Place-Based SameTime Awareness Place-based awareness facilitates useful discussions Instant messaging Instant teamrooms Membership
The Knowledge Discovery Server Connects people with the right info at the right time – Integrates People, Places, Things into a Knowledge Map – Discovers relationships between People, Content and Categories to add context to information Supports KM practices within organizations – Respects user privacy – Enforces system security
What Does the Discovery Server Do ? Out of the box Discovery Server will: – create a knowledge map – generate affinities – create expertise profiles – assign content value – index everything – cluster and organizes documents – relationships b/t people and topics – mine skills, locate experts – based upon computed metrics – search for docs, people, topics, etc. n Discovery Server components constantly maintain and update themselves through a combination of automatic processes and administrative tools
Discovery Server K-map User Interface
A Vendor-neutral KM architecture
Mapping KDS to the Architecture Discovery Server Knowledge Map Browsable/Searchable Topic map of People, Places, and Documents Solutions Application templates + Methodologies + Services K-station Portal Organize and manage personal and community assets Metrics
Content Spiders: – Lotus Notes/Domino, Domno.doc, QuickPlace, Filesystem, Web (HTML), Directory Spiders: – LDAP Server V2 or V3, Domino Directory/databases Spider: Notes Enterprise Data Spiders: – Domino/Notes Spider with DECS and Lotus Connectors – Content Spider SDK Content stores: Spiders
People/Partners B2B ebus Applications Data Legacy Structured Unstructured Enterprise Type text Enterprise Commercial & External Feeds "Write only" memory Unshared tacit knowledge "Write only" memory Unshared tacit knowledge TRADITIONAL ENTERPRISE "Write only" memory Unshared tacit knowledge RELATIONSHIPS - ACTION "Write only" memory Unshared tacit knowledge People related to Content Content hierarchy Categories Expertise Information warehouse: Content Catalog
Search Content Valuation Search Expertise Search Hot Lists Information Warehouse: The K- map Search Catalog Metrics Search Communities Search Content Valuation Search Expertise K-map
Pets & Animals Veterinary Help Clubs & Associations Plants & Ponds Aquariums Aquarium Keeping Fish & Livestock Traveling with Pets Health & Vet Help Cats Birds Unusual Pet Animals Products & Services Saltwater Fish Zoos & Aquariums Horses Fish & Aquariums Advice & Guides Category Labeling - Applies a human readable tag to a category Clustering/Categorization - creates categories of similar documents and moves new documents into the appropriate categories (IBM Research technologies) Classification: K-map Builder
Pets & Animals Saltwater Fish Veterinary Help Clubs & Associations Plants & Ponds Aquariums Aquarium Keeping Fish & Livestock Traveling with Pets Health & Vet Help Cats Birds Unusual Pet Animals Products & Services Zoos & Aquariums Horses Fish & Aquariums Advice & Guides Kmap Editor - Manages relationships between documents and categories Affinities - Matches people with categories based on their interaction with the documents in the categories Metrics - Calculates value of documents and strength of affinities based on use Classification: People
Retrieval: K-map Indexer Search content across the information warehouse Scope your searches, find only what you need – Everything About – Documents About – Documents Authored By – People Named – People Who Know About – People Whose Profile Contains – Places About – Categories About
Presentation: K-station Portal with common structure Create shared places from templates Put information in context Reuse places as templates
Presentation: K-station portlets
Association: Metrics Metrics – Collects "Digital Breadcrumbs" Statistics about information flow – No additional burden on users – Leverages document meta data – Analyses trends, relationships, and patterns
Authorship - documents created by person Linkage - number of links to/from a document Messages - number of messages between two people, number of links forwarded Activity of document or database - frequency of change, volume of change Activity of person - frequency of system use Association: Basic Metrics
Association: Advanced Metrics Advanced Metrics are calculated using basic metrics and relationships between entities Person to Topic Affinity - based on documents in the topic and the people who authored, contributed, distributed, and read Value of Document - based on activity of document, linkage Value of Topic - sum of Value of Documents in Topic
Reports Content and Usage Activity: – Most Active K-map Categories – Highest Document Values – Most Active Authors – Most Active/Linked To/From Documents – Most Active/Read Documents Monitors Activity Trends over Time Association: Metrics Reports
Tales from the trenches Set appropriate expectations – Map to a known business process – Determine access to content stores in advance – Anticipate some effort to create and maintain the taxonomy – Look at existing meta-data If creating user profiles and affinities, consider privacy issues
KM Discussion Lotus KM Product Information