Project Intro: Research Knowledge Manager

wrendeceitInternet και Εφαρμογές Web

21 Οκτ 2013 (πριν από 3 χρόνια και 7 μήνες)

76 εμφανίσεις

©2013 ADS Alliance Data Systems, Inc.

Confidential and Proprietary

Project Intro: Research Knowledge Manager

1

Project Intro: Research Knowledge Manager

Jeff Buchmiller

Business Research & Intelligence Group

Alliance Data

August 2013

jeff.buchmiller@alliancedata.com

214
-
494
-
3431

©2013 ADS Alliance Data Systems, Inc.

Confidential and Proprietary

Project Intro: Research Knowledge Manager

2


Customers get more



products they want


Customers save money


Company performs better

Company: Alliance Data

©2013 ADS Alliance Data Systems, Inc.

Confidential and Proprietary

Project Intro: Research Knowledge Manager

3

Function: Business Research & Intelligence

Some of My Job’s Goals

1.
Enhance the Business Understanding model to cover how the contributions of our people and the
investments we make in them impact the enterprise’s outcomes


Need a layer of abstraction to organize the knowledge represented in 1,000’s of documents and reports


My own special focus is on the workforce and organizational productivity aspects; each of my teammates
has a different focus (economics, products & services, and marketing strategy), and we share the rest


2.
Produce an Alliance Data "Workforce State of the Union" report


The prototypical example for how the project’s deliverable will be used


manages the data, information, &
knowledge when it is too large to be held in my own head, helping me to pull from it quickly and easily


3.
Provide research & analysis subject
-
matter expertise and education


Same as above, except very specific requests are handled on an
ad hoc

basis


©2013 ADS Alliance Data Systems, Inc.

Confidential and Proprietary

Project Intro: Research Knowledge Manager

4

Project: Research Knowledge Manager

Repository of Knowledge


Documents


Reports from internal and external sources


News articles from the web


Miscellaneous data files


Videos


Images such as webinar screen shots


Email attachments



I myself have accumulated over 31,000
documents in the past few years


©2013 ADS Alliance Data Systems, Inc.

Confidential and Proprietary

Project Intro: Research Knowledge Manager

5

Project: Research Knowledge Manager

Purpose of the Project


I use Google Desktop or Microsoft Indexing Service to index all the documents collected, then
perform keyword searches to locate information needed to answer the question at hand.



It takes quite a while to locate all the relevant information, now that the document library has
grown so large.


A significant portion of the time comes in having to open and read each document that my
keyword searches “hit,” to evaluate whether it really does contain useful information.


Another significant consumption of time occurs because the number of documents that my
searches “hit” is often large.



We need a better knowledge management tool that can help us get to the right information more
quickly and reliably.

©2013 ADS Alliance Data Systems, Inc.

Confidential and Proprietary

Project Intro: Research Knowledge Manager

6

Project: Research Knowledge Manager

Software Features


The deliverable is to be some form of desktop software that will index, tag, and search, at a
minimum, the content of the documents.


There is just one user per installed instance on a desktop or laptop PC.


There needs to be a user interface for adding documents, tags, and other information into the
system, and a user interface for performing searches.


We use Windows XP, Windows 7, or Windows 8.


Indexing could be accomplished similarly to existing technologies. There are different kinds of
indexing available, including more semantic versions that “understand” the meaning of content
and queries better than keyword indexing even attempts. There is a great opportunity to adopt
and apply natural language processing and semantic web techniques to this problem.


Some conscious tagging by the user of content within documents can be extremely useful for
later retrieval.


For example, there may be certain statements in documents that are related to “Big Data”
and these can be tagged by the user, so that great search results specifically for the
popular topic of Big Data are enabled, which may be impossible for any keyword search to
accomplish.