Magazine article Online

Datastream: Numeric Data - All You Can Use at a Fixed Price

Magazine article Online

Datastream: Numeric Data - All You Can Use at a Fixed Price

Article excerpt

Numeric databases are different from bibliographic databases. Yes, they have more numbers, but they also have the following characteristics that make their use by end-users in libraries difficult:

* Specialized Systems: Most numeric databases are on systems that are unfamiliar to those of us who were brought up searching DIALOG and BRS. The bulk of numeric databases in finance and economics reside on systems such as DRI, Wharton Econometrics, and LP. Sharp. Our unfamiliarity with the systems inhibits their use, since we must first learn a new search protocol and then instruct end-users.

* Approach to searching: Patricia Suozzi points out that we don't "search" numeric databanks, we "retrieve" from them. We determine what numeric series the databank has and use the appropriate code to retrieve the data. Both the relevance and recall of our retrieval should be 100%.

* Online Manipulation: Numeric databases often allow the user to "massage" the data online. Data manipulation can take many forms including creating new variables, statistical analysis, and displaying the data as graphs. As with the search protocols, the library staff must first learn the techniques of data manipulation before instructing end-users.

* Cost: Time-sharing numeric databanks usually charge for CPUs (computer processing units) as well as for connect time and data items printed.

For a library, the major barrier to providing direct end-user access to specialized numeric time-sharing systems is their high and unpredictable cost. Online bibliographic business data is certainly not cheap, but bibliographic time-sharing systems often give libraries inexpensive alternatives such as after hours searching, lower subscription prices, and academic discounts. In addition, bibliographic data is increasingly available on CD-ROMs for a fixed cost.


A time-sharing system that makes end-user searching of numeric data a real possibility for the library is Datastream. Datastream is a UK-based information utility that is a subsidiary of Dun & Bradstreet. Access to Datastream is by annual subscription. The annual fee buys one password that allows unlimited access to the system, as well as documentation, training, and user support services. There are no extra costs.

Datastream specializes in international financial and economic statistics. A partial list of the types of data the system contains includes:

*Company Accounts: Financial accounts for 3,000 companies in eight countries.

* Economics: 35,000 monthly to annual series from the OECD, the IMF and the Bundesbank. The time series begin as early as 1950.

* Exchange Rates: Daily rates for 80 currencies. Detail includes spot, middle, forward, and trade-weighted rates. The series begin as early as 1965.

* Fixed Income: Contains details on 45,000 bonds from 13 countries. The series is updated daily and begins as early as 1969.

* Futures: Financial and commodity futures from 19 exchanges in Europe and North America. The frequency is real-time and series begin as early as 1977.

* Market Data: Daily stock market data for 9,000 stocks in 30 countries. Most of the series begin in 1973.

* Stock market Indices: 600 indices from 28 countries beginning as early as 1965.


Datastream's front end software simplifies the process of searching. Mnemonic codes for variables can be looked up online or copied from printed code books. Elaborate help screens guide the user to the appropriate program and codes. Searching the system is a two-step process. The first step is choosing the appropriate program. The programs are similar to separate software packages. There are individual programs for the display of time series, company accounts, for graphics, and for statistical analysis. The user retrieves data by filling out a template. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.