By Tawheeda Wahabzada and Chandrika Kaul
Earlier this year, Open Data Watch (ODW) released results from the 2015 Open Data Inventory (ODIN), a comprehensive assessment of the coverage and openness of data available from 125 national statistical offices (NSOs). Conducting an ODIN assessment requires hours of intensive work searching for data on NSO websites. During their work, ODW assessors recorded their observations on the functionality of the websites, noting characteristics such as the logical organization of data; relevance and viability of web links (broken links and error messages); inclusion of contact email/phone and other information to communicate with the webmaster or department responsible for data; relevance of graphics that do not distract from web content; inclusion of information about the organization and its mission statement; and the currency and regularity of updating information. They found both good and bad websites. From their experience we have distilled a few pertinent lessons.
♦ Logical organization of information is critical for efficient NSO website navigation. The organization and contents of a website reflect the engagement of the NSO with their users and their commitment to providing a public service for all. With proper data arrangement, users can easily locate data by topic. For example, the Republic of Macedonia State Statistical Office’s database is cleanly and logically organized. The data are arranged in folders and subfolders with clear subject labels. Along with a listing of the contents of its databases, ODW recommends that NSOs provide a Frequently Asked Questions (FAQ) or Help file to guide users.
♦ Administrative or technical barriers exist when users cannot access data without “subscribing,” providing login information, or downloading special software. These may discourage users who find the requirements burdensome or who don’t wish to leave a record of their transaction. The Department of Statistics of Malaysia website does not allow users to access any free publications without logging in. This additional step hinders users from accessing otherwise free and unrestricted publications. To further improve the accessibility of data, NSO websites may provide a data request form for users who want to obtain data that are not provided in the NSO website. Here is an example of a data request form from the Statistical Institute of Jamaica.
Figure 1: Example of a CC BY icon, National Institute of Statistics of Rwanda
♦ Metadata availability – According to the United Nation’s Guidelines for the Modeling of Statistical Data and Metadata, metadata “provides information on data – and about processes of producing and using data.” Providing data about data is essential to ensure credibility of the data and to assist users in understanding data. Ideally, a NSO website should have a webpage on metadata that is as clearly organized as its data pages. The State Statistical Committee of the Republic of Azerbaijan provides such a webpage, which is easily accessible from the home page. Metadata should also be downloadable with data.
♦ Search capacity – As NSO websites expand by adding more data, the need for a search function becomes essential. A search box removes barriers to data accessibility for users and researchers. Based on the assessment of 125 NSO websites countries, more than 75 percent include a search box. Mexico’s Instituto Nacional de Estadistica y Geografia provides an effective and functional search box. By simply searching for “PIB” (GDP), the results page provides over 200 results. Despite the sizeable search results, the results page provides “suggested results” and “searches related to GDP” (as seen on figure 2) to help users refine their search.
Figure 2: Example of search results, Instituto Nacional de Estadistica y Geografia Mexico
♦ Regular maintenance – An NSO website should provide up-to-date statistics. It must also be reliable. The webmaster should ensure that the links throughout the website are not broken or do not lead to 404 errors, indicating that the webpage is not found. Additionally, server availability is critical to providing excellent user experience. ODW assessors experienced occasional server outages, where some NSO websites were down for several weeks.
Figure 3 displays a collage of NSOs lacking regular maintenance. The top left image is an example of a 404 error in the contact information webpage of Burkina Faso’s NSO. The top right image is an example of server unavailability of the State Statistics Committee of Turkmenistan website. The NSO website was disabled for over two weeks, yet as of April 26, the website was no longer disabled. The final two images are examples of the Central Statistical Organization (CSO) of Yemen. The homepage of the NSO website displays a “deceptive website” warning. Even after bypassing the warning message, the NSO website is still inaccessible, as it is suspended.
Figure 3 – Examples of NSOs lacking regular maintenance
♦ Page-loading time is important for a data heavy NSO websites. Some NSO websites were slow at loading PDF reports, posing challenges on the accessibility of data (the rule of thumb for loading reports is 20 seconds). Additionally, several websites required unzipping large files, which in turn requires users to have an unzip utility installed on their computers.
♦ Language considerations – ODW encountered a few NSO websites that had inconsistencies in languages. A good example is of India’s Ministry of Statistics and Program Implementation, where efforts are being made to have government websites in both English and Hindi. However, in the Hindi version of the website, only the homepage is in Hindi. The hyperlinks provided on the homepage of the Hindi version lead to English language publications and other webpages. If applicable, NSO websites should maintain language consistency and provide the same content in all languages.
A good website is ultimately the result of a strategic combination of people, processes, and technology. Around the world, national governments and their statistical offices are working to improve the quality and availability of important statistical information. New technologies are making it easier to collect and disseminate data, and the demand for data is growing. The NSO’s website is the vital connection between data producers and data users. There is no single, correct design. What is essential is that it provides open access to the most reliable data to the widest possible community of data users.