Opening the government’s data kimono

This year, the Hong Kong government plans to open up over 650 datasets to be released online for the public to use and view freely. It sounds great, but some experienced legislators and tech experts know the devil is in the details.

The new datasets cover different areas, including data related to the economy and livelihood, real-time meteorological data, geospatial data, digital maps that can facilitate smart city development, and data that can enhance city management.

“By the end of 2019, the number of datasets on the [data.gov.hk portal] will be increased from around 3,300 to close to 4,000 in total, an increase of about 20 percent,” says Mr Victor Lam, the Government Chief Information Officer.

Last week, the Office of the Government Chief Information Officer said over 80 government bureaux and departments already published their first annual open data plans, including the Environmental Protection Department and the Food and Environmental Hygiene Department.

According to the government, the datasets will be released in machine-readable formats commonly used by the industry, such as JSON, XML and CSV. The public can browse, download, distribute, reproduce, print and hyperlink all datasets disseminated via the website for free, for both commercial and non-commercial uses.

This responds to the city’s ambition to turn itself into an innovation and technology hub. The government says one approach is to open up government data to provide raw materials for technology research, innovation and smart city development.

Bring it on

IT legislator Mr Charles Mok welcomes the plan, saying it is significant for researchers and developers.

“Since the launch of the government’s open data portal in 2011, it is the first time that the government commits to releasing annual open data plan for all bureaux and departments,” he says.

However, Mr Mok points out that making more datasets available to the public is not enough. He calls for an open data policy that emphasizes data quality and meets the public needs.

Raw data is preferable to aggregated data of website content or statistics to ensure good quality, and datasets should be as up-to-date as possible.

Mr Mok says that datasets that are higher in demand, such as data related to public health, can be published in API in real-time or of higher frequency.

Room to hide

The plan also does not touch on the criteria to determine the priority of data releasing by different government units.

“Government departments have much flexibility to decide which, how and when to open up datasets. Information that is of high public interest and concern, such as government expenditures, declare and disclosure of interests of the government officials, real-time transport data and so on, appears to still be lacking,” he explains.

“They should set a schedule and deadline of delivering the ‘open by default’ objective, and report to the public regularly on the public’s requests for data and how much of those are met. Accountability and transparency should be ensured,” he adds.

The current open data policy also fails to include public institutions and quasi-government bodies, as well as the private sector.

PPP based on trust

Hong Kong may learn from the UK’s “data trust” arrangement to incentivize the private sector to share their data to provide the public with real-time transport information.

The UK’s data trust model involves a mutual organisation being formed to manage data on its members’ behalf. Data subjects would pool their data forming a trust, stipulating conditions under which data could be shared.