EDS Program Study

  EDS > Program Study
printer friendly version Print
Page

Table of Contents
Introduction

The Electronic Data Service (EDS) is operated jointly by the Columbia University Libraries and Academic Information Systems (AcIS) to support instruction and research that involve numeric data resources. The following is a program document that describes the collections and services to be provided by EDS in a new facility, IAB 215. The plan will consider the trends EDS staff have identified in both the EDS collections and services in order to greatly enhance the provision of existing services as well as anticipate the needs of EDS in the future.

The assessment of our future needs was done by first examining the general trends in data services and then by reviewing developments in our current operation (collections, software support, consulting services). We've also included detailed information about GIS services, which we would like to better serve with the new resources available in IAB 215. We've provided some information on other GIS service operations based on interviews with libraries and data centers providing GIS support.

A summary of the resulting basic requirements for an EDS facility is listed here, followed by the full report which details the findings from our assessment process.

Public space should provide for:

Staff space should provide for:

General Trends in Data Service

Two documents were used to examine general trends for how data services are being provided at other academic institutions. The first is the Association for Research Libraries Spec Kit 263, Numeric Data Products and Services. It summarizes the results of a survey of 66 institutions conducted in August 2001. EDS provides most of the services mentioned in the survey, with the exception of data analysis and instruction, each of which was provided by only 31% of respondents. In terms of data formats, memberships in consortia, and types of data providers, EDS used all the alternatives that were reported as being used by at least one respondent.

EDS does not assist with statistical analysis but does specialize in helping users with the tasks of subsetting, merging, and converting between file formats; tasks we consider part of the retrieval process but tasks that were not mentioned in the report. The terms used in the survey about assisting users "to retrieve data" and do "data analysis" are not clearly defined so the counts for those doing "data analysis" may be overstated based on how EDS would report its activities.

EDS does not do formal instruction beyond class presentations about our services, and in its preparation of user guides covering both the use of statistical and GIS software and the use of specific data products. EDS has the expertise and interest in providing basic instruction for the software packages we support but we do not have the space and equipment to do this.

The second document was "Data Basics: An Introductory Text", a booklet prepared for the August 2002 ICPSR workshop, Providing Social Sciences Data Services. It gives a much more in-depth description of the range of tasks that can be included within a data service operation, and the document separates the services of "identifying", "extracting", "data analysis," and "delivery, copying, and subsetting". Again, with the exception of data analysis, EDS provides most types of service. In fact, EDS was used as an example of one model for service at the workshop.

The booklet also describes "statistical reference" as a task that may or may not be included when defining a data service model. Statistical reference is presented as being different from traditional data service and one that should be viewed as a potential addition to traditional services rather than a substitute. They make the point that, with the trend to publish statistical tables on the Internet or on CD-ROM/DVD rather than in print, the distinction between users of statistics and users of data is blurring.

The impact of that blurring on service mentioned is occurring here at Columbia. As recently as five years ago when most statistical tables were published in print, someone coming to EDS seeking such information would have been referred to the Lehman Reference desk. Today it is likely for the reverse to happen because statistical tables published on the Internet and on CD-ROM/DVD are easily accessible to users and their electronic format serves as an added bonus. This means the users can simply import the tables into their papers, spreadsheets, or even statistical software, providing them with an excellent time saving device. Additionally, the details within some of these tables have grown beyond what was possible in a print publication, with some sites being interactive allowing users to generate custom tables.

Statistical reference represents a significant portion of the walk-in traffic in EDS. The mix of statistical products has been changing with some switching to the Internet while other popular ones continue to be published on disc. This is because these products come with powerful software that web interfaces cannot match. The Census 2000 Summary File 3 product, the most popular of all Census products, has just been released. The DVD version of this file will be not only the most powerful way to access this data, but the simplest. EDS expects a surge in demand when it becomes available, and will be prepared to demonstrate to statistical users how to use the product to retrieve a few rows of numbers as well as prepared to advise serious users as to when it is appropriate to use the DVD version as opposed to the raw data files available in EDS or via the Internet.

Collections

EDS's core function is to provide access to datasets in a number of ways, including the maintenance of a large library of data files and their relevant documentation. The current inventory includes data files from well over a thousand studies, comprising thousands of data files. Our collections are mostly electronic and will continue to grow as our major providers expand their holdings. The general trends are:

These trends can be seen by looking at our major data providers.

Services

As a result of combining the AcIS statistical package expertise with the library's data access services, EDS Services include data consulting, data access, and instructional support beyond the acquisition and storage of data files. Data consulting includes advising users on the extraction, merging, converting, transferring, and compressing of data. Our service model empowers users to perform tasks themselves whenever feasible, which can happen in many different ways.

When a user enters EDS, the consultant firsts ask the user a series of questions we refer to as the data interview, which helps us ascertain which product is appropriate to the user. In some cases, the user knows right away which dataset they've come in to access, and we are able to get them started very quickly. If they've never used the software which accesses the before, we show them how to open the program they need and either provide them with an EDS User Guide which explains how to use the product step-by-step, or we provide them with a few verbal cues in order for them to begin. In this case the consultant then will go back to their other work, allowing the user to work independently until he or she has another question.

Users that do not already have in mind what it is they are looking for require a much more comprehensive data interview, so that the consultant can understand what research question the user needs to answer in order to help them locate applicable data. When neither the consultant nor the user know what data would be appropriate, the two will generally begin searching DataGate and our other Internet catalogs. If nothing appears fruitful, the consultant and user will browse the web together in an attempt to turn up more leads.

Tasks such as converting, transferring, or compressing data may also require greater levels of assistance. Again, whenever possible, the user performs these tasks on their own under the direct supervision of the consultant. The users are responsible for performing all the steps on the computer themselves, although in some instances the consultant may demonstrate a step for the user to help the user understand what he or she is about to do. We find that many users do not require our direct assistance on return visits to EDS due to this thorough orientation to the processes required for these operations.

We advise on five major categories of software:

Point of Service

Currently, EDS patrons have the following point of service options to choose from when requesting EDS services:

Of those options, the majority of assistance with EDS services serves walk-in traffic.

Because Columbia affiliates have greater and greater access to computers and, consequently, electronic mail, it is important to gauge whether or not EDS has been experiencing a decline in foot traffic. In examining the EDS tracking data from 9/23/99 to 8/23/02, we find that EDS did in fact, experience a decline in walk-in use of EDS between the fall of 1999 through 2001. However, that number has picked up again in 2002 to the rate that it was in the fall of 1999, and we have reason to believe that trend will continue as we provide greater support for statistical packages and begin support for GIS assistance.

Table 2: EDS Point of Service Entry Types by Year
(Partial Year Data for Both 1999 and 2002)
Year Appointments Electronic Mail Telephone Walk-in
1999 - 07.41% (16) 01.39% (3) 89.35% (193)
2000 00.24% (3) 14.48% (180) 02.49% (31) 80.13% (996)
2001 00.06% (1) 15.81% (247) 02.75% (43) 77.40% (1209)
2002 - 07.44% (63) 02.48% (12) 88.67% (751)

Consultations in EDS vary in length of time (see table below for reference statistics). Most commonly, users make contact with the consultants when they first enter EDS to gain some assistance getting started with their work. They then begin working on their own, asking consultants for further help as the need arises throughout their session. For example, a user may come in and sit down to work for a two hour period, perhaps asking three or four short questions of the consultant. This type of interaction is reflected in our relatively higher number of less than fifteen minute consultations, as the same user is sometimes recorded several times over a single visit to EDS.

Alternatively, users may come in for intensive consultations that may last several hours. Moreover, users requiring lengthy consultations often require several consultations depending on what kind of task it is they're attempting, which may happen over a week or sometimes a more drawn out period of time.

Table 3: EDS Reference Stats For Walk-in Transactions
Week <15 Minutes 15-30 Minutes >30 Minutes
02/11/02 45 8 -
03/25/02 63 12 5
04/15/02 62 14 3
06/10/02 15 5 1
Staffing

Staffing is provided jointly by the Libraries and AcIS. None of the EDS staff work there full time. The data librarian oversees the day-to-day operation, manages the collection development, and provides reference service. There are three AcIS staff members who each provide general software consulting plus the following: one serves as the GIS specialist and maintains the software on the network, one provides SPSS consulting and maintains the Cunix-based DataGate catalog, and one serves as the SAS consultant. Two other AcIS members contribute to the tasks of long term planning and network design but are not involved in the daily routine. Graduate students complete the staffing, and their assistance enables us to ensure that there are at least staff two persons in EDS during the hours it is open. The number of graduate students on staff varies between two and four depending on the time of year.

The data librarian has a desk and computer in EDS. All other staff share a desk and computer. This computer is also used for consultations that involve special software installation or writing to a CD-ROM. With the move to IAB 215 plans should include the addition of a workstation for AcIS staff and one for a GIS/map librarian.

Space and Equipment

Flooring and Furniture Layout
Both the open staff area and the public areas in 215 are carpeted. The carpet is shabby, worn bare in spots, stained, and fraying along its seams. It is glued to the floor. Two metal power strips extend out from the north wall to at least two thirds of the way across the room. These strips now sit under two long rows of closely packed computer tables. This configuration will need to be changed, as EDS workstations need to accommodate more than one user. These strips will have to be removed. Once they are moved, the carpet will have a large gap, making it dangerous for people walking around.

The north wall is better suited for shelving and seating for using the manuals and consultations rather than workstations since there is electro-magnetic interference coming from beyond the north wall. It currently causes flickering on the screens of the AcIS computers. The east and south walls are better locations for workstations.

In terms of furniture, we are fortunate to be able to use some of the existing furniture in IAB 215. Because we wish to provide users with workstations which accommodate multiple users or users sitting down with consultants, each workstation will consist of two tables and two chairs, but only one computer. We will also continue to provide a few empty tables with chairs for laptop users. Any remaining tables will be used in the staff areas as desks and work tables.

Our recommendation is that in the public area the carpet be removed and tile put down. Because the flooring under the carpet in the staff area is probably wood and not concrete like under the carpet in the pubic area, we are not asking that the carpet in the staff area be changed. Because the expenses for furniture will be low, there should be room in the budget for the cost of this improvement.

Finally, we will need to have both a coat closet for staff working in EDS and a supply shelf located in the staff area. We should be able to use furniture from our current facility to meet those needs.

Shelving
Based on measurements of our existing shelving which includes print codebooks, manuals, documentation, and CD-ROM booklets, we will need nine single-sided wall units of library shelving, which includes room for anticipated growth. The units are 36"w x 9"d x 84"h and will be set up for 6 shelves each with canopies.

We plan on using available shelving from Lehman. It will need to be determined if all the parts required are available in Lehman, or if we will need to order some pieces. Also, we plan on removing and re-assembling 2 units which are currently in 312 Lehman. This should be included in the expense for constructing the shelving in EDS.

Hot Slots
We require wall mounted hot slots with capacity to house at least forty EDS User Guides. Currently, we have only one wall mounted unit which stores only 10 guides.

CD-ROM Storage
We have an 8 drawer CD cabinet that has 840 slots and is approximately 80% full. With the empty space spread throughout the portion of the collection that is growing, growth space is quite tight. We also have over 300 CDs that are stored elsewhere. The trade collection, which is not in the cabinet, is growing at a rate of 2 CDs per month. EDS needs one more CD cabinet to store our collection.

Computers and Related Equipment
Under our current operation, we maintain 11 computers and also house one AcIS lab machine. Of those 11, nine are physically located in EDS. Seven of those are public workstations, one is reserved for staff to install special datasets and their software on an ad hoc basis upon user request, and one is for the use of the data librarian. The remaining two machines are located in Watson Hall. The first computer serves as a way for AcIS staff to provide remote support to EDS from Watson Hall, while the second is for development purposes only.

These machines are nearing the end of their life cycle. They will need to be replaced with machines which feature GIS capability, large flat-screen monitors (17" viewable), and CD writing capability. Additionally, we require an additional three public workstations and two additional staff machines to accommodate an increase in traffic to EDS with adding GIS support to our repertoire of services. Of the two additional staff machines, one would be set aside for the GIS/map librarian. Because equipment in EDS is standardized to provide optimal performance for our users, it is essential these additional computers are included when our current machines are necessarily replaced. We are not recommending replacing our EDS server at this time.

In addition to the computers, EDS currently has its own black-and-white laserjet printer which is used by EDS staff to print codebooks and documentation as well as by patrons. We would like to include with our GIS/map services a large format printer or plotter and a large format scanner. The availability of a scanner is certain to be an invaluable asset in Lehman library, and would help integrate new GIS services with the existing map collection. Issues of cost recovery for large scale printing and scanning are open for discussion.

[an error occurred while processing this directive] Thursday, 07-Oct-2004 15:24:59 EDT [an error occurred while processing this directive]