“Design, query, and evaluate information retrieval systems.”
Introduction
Tucker (2023) describes how designing an information retrieval system is an iterative process. The database designers must consider what the system needs to do, what kind of data it needs to handle, what it should look like to the user, and plan how to evaluate whether the system is working as intended. Users of the system should be asked for feedback, and multiple stakeholders should be involved in the design process. Changes should be made based on feedback and evaluation. Considerations should be made for data growth and new technologies, however, database designers should be wary of adding features just because they are new and should instead consider whether a new feature fits into the needs of the system.
Information Retrieval System Design
The design of an information retrieval system has a large impact on whether a user will be able to achieve the desired result. For example, if a user needs to be able to search for items by their creation date, the database needs to have a field for creation date attached to each item. Clear guidelines must be written to make sure that the definition of “creation date” in the mind of the end user and the definition of “creation date” in the mind of the person (or computer) putting the data into the system are the same.
This is fairly simple when working with numerical data such as dates, because they are easy to define, but when words are used to describe an attribute or feature of an information item, it can get complicated. This is why we need controlled vocabularies. A controlled vocabulary is a list of what words are appropriate to use in the database. For example, should I write “rabbit,” “bunny,” or “hare” when describing a specific long-eared mammal? It should be clear to both the indexer and the searcher what terms should be used. Words with multiple meanings should also be addressed in a controlled vocabulary. If a user types in “ring,” can the system determine if they mean a piece of jewelry, a sound, or a household security system? Good databases have controlled vocabularies that account for these issues. Many controlled vocabularies coordinate words together to help the user understand the terms, and some allow the user to coordinate words. For example, combining the word “ring” with “security system” or “jewelry” would help the user find the specific concept they are looking for.
Querying or Searching
Although many people are used to going to a search engine such as Google, typing in a few words, and getting a basic answer to a question, Tucker (2014) describes how “expert searchers” can search more effectively and find what most people overlook. Expert searchers are more aware of the quality and quantity of information that can be found and take a more active role in searching for the information they need. I have seen this in myself as my searching abilities have grown. In the past, I would go to the library, survey the books on a shelf, and select something that looked interesting. Now I look in various online catalogs and book review websites for a book on a specific topic until I find what I am looking for. Often that book is not available at my small local library and I have to order it from another library or purchase it, but I am ultimately satisfied I have found the book I am looking for.
Knowledge of how a database works is essential to finding what I am looking for in a database. If I search a string of words, will the search engine look in the title of a document, the body of a document, or the metadata used to describe the document? Do I need to search using a specific controlled vocabulary, or will the search engine try to match my search terms to the controlled vocabulary? What kind of content is indexed in this database? Do I need to look for a different database to search for the type of material that I want?
Many search engines have ways to search for a specific phrase, and to indicate whether two terms should both be in the search results or whether only one of them is needed. Other common features include a wildcard or placeholder that will help you search for different spellings or uses of a word (e.g. music* could include music, musical, musician, etc.) and excluding certain words or phrases from the search results. Different search engines will have different ways of implementing such features, so it is important to become familiar with the specific search engine you are using before running a search in order to ensure that nothing is overlooked.
Evaluation
To judge whether an information retrieval system is achieving its desired goals, we need to know whether users are finding what they are looking for or not. Even if exactly what the user wants is not in the database, the user should be able to see a decent number of search results that are relevant to their search without much trouble.
User testing can also provide feedback on the layout of the website or database software. Krug (2014) discusses how users want to be able to find what they are looking for fast. They should be able to find the button they need without much effort. Both good interfaces and good search databases are designed with the idea that the user wants to do as little work as possible.
Items of Evidence
Evidence Item 1: Online Tea Shop Group Project and Reflection
My first item of evidence is in two parts. The first is a paper produced as a group project by me and two other students for the course INFO 202: Information Retrieval System Design. The second is a reflection essay with my thoughts on the project and my role in the group. As a group, we created a database for a hypothetical online tea shop. The paper outlines the use case and expected users. It also lists the data entry fields and the rules for entering data into those fields, as well as screenshots of the search page and submission form of the database.
This item shows I have experience designing an information retrieval system with a team. It also shows that I understand how to conduct user research and tailor such a database to a specific user group.
Evidence Item 2: American Film Institute Internship Overview
My next item of evidence is a screen recording video where I walk through the steps I used to conduct research for my internship at the American Film Institute. During the internship, I gathered information from historical entertainment industry magazines, newspapers, and other sources to enter into the online database of American films published on the American Film Institute Website.
This shows my understanding of different methods for querying databases and my ability to discern whether or not a particular database will be useful for conducting a specific type of research.
Evidence Item 3: Evaluating Different Databases
This piece of evidence is a discussion post I wrote for the course INFO 202: Information Retrieval System Design. In the post, I evaluate three different types of databases: Google Scholar, a library database, and a retail website. This item shows my ability to analyze different types of databases and how they are targeted to different user groups.
Evidence Item 4: Website Navigation Redesign Group Project
My last piece of evidence is a group project paper I made with three classmates for the course INFO 202: Information Retrieval System Design. In the paper, we discuss the retail site Life Without Plastic, and how the site and its retail database could be redesigned to be easier for target users to navigate. This shows my ability to evaluate and redesign an information retrieval system with a team. I came up with the idea for which website to use and typed up an outline for the redesign plan based on discussions with my group. I also created the graphics that illustrate the current site map and the proposed redesign.
Conclusion
Information retrieval systems are a key part of the information profession, and one I find quite fascinating. They are a part of all of our daily lives, and a deeper understanding of how they are designed and how they function will help me be better at storing and retrieving information for my own use and making it available to others. This aspect of information science is heavily affected by technology, which changes rapidly. To stay up to date in this field, I will follow publications from organizations such as the Association for Information Science and Technology (ASSIS&T) and seek out opportunities for continuing education to learn about new technologies.
References
Krug, S. (2014). Don’t make me think, revisited: A common sense approach to Web usability ([3rd ed.]). New Riders.
Tucker, V. M. (2014). The expert searcher’s experience of information. In C.S. Bruce, K. Davis, H. Hughes, H. Partridge, & I. Stoodley (Eds.), Information Experience: Approaches to Theory & Practice, (pp. 239-255). Bingley, UK: Emerald Group.
Tucker, V. M. (2023). Information retrieval system design: Principles and Practice ([6.2 ed.]). AcademicPub/XanEdu.