How can governments share data more easily?
By Informatica
A concept from Netflix can help.
Data governance “gives you a pulse” on how usable your data is, says Anand Ramamoorthy, Director, APJ Head of Data Governance and Data Security at software and data firm Informatica. Managing data well enables organisations to extract value from their data and share it more easily.
Good data is the bedrock of more advanced tech like analytics and AI. Anand explains how organisations can build a firm data foundation to serve citizens better.
The problem with data
In an ideal world, all datasets would be complete, accurate and secure. But that’s rarely the case. Anand gives four issues government datasets often run into.
The first is incomplete data. Governments can only design policies based on the information they have. If a date of birth dataset is only 70 per cent complete, any policy the nation implements for the aging population may not be very effective.
Once governments know that a certain dataset is lacking, they can then find ways to fix the problem. Perhaps date of birth wasn’t a mandatory field in some government forms. Agencies can change that to make the dataset more complete.
The second problem is duplicated data. Not knowing what datasets each ministry has might lead to duplicate sets of data sitting in different parts of government. For instance, the health agency might collect patients’ financial information to know if they qualify for government subsidies, instead of approaching the finance ministry. This is risky, as more datasets means more opportunities for leaks or hacks.
Third, organisations may not fully understand the value of their data. With the varied sources of formats of information available today, governments need to understand what types of useful data they can collect from different sources.
For instance, social media is a good way to understand public sentiment, Anand says. “Are people saying they need to wait a long time at hospitals? Are they not getting what they want?” Public health agencies can gather this information and apply analytics on top of that.
Fourth, the source of a dataset may not always be clear. But this is crucial information for sharing data across and within agencies. “Otherwise it’s just the blind leading the blind,” he says.
What can help
An underlying problem is the lack of visibility to data. Organisations need to know where the gaps are before they can address them.
This is where good data management can help. Informatica’s data governance tools use automation and AI to help organisations understand how usable their data is.
Informatica also allows governments to pool their data in a centralised data portal. “This is a marketplace format, which holds data in a much more reusable manner, rather than in an Excel sheet,” Anand explains.
Agencies would be able to see the metadata of each dataset in this portal. This makes data sharing easier, as agencies can tell others what types of data they have without exposing it.
Metadata would give the necessary information to help agencies decide if a particular dataset would be useful to them. For instance, they could see the accuracy and completeness of the dataset. Agencies can decide whether these datasets are good enough for them to work with.
It’s just like scrolling through Netflix, notes Anand. “You’re able to see the director and ratings of the film, and this gives you enough information to decide whether you want to watch it,” he says.
If agencies do want a dataset, they would have to ask for permission. The agency that owns the data can clarify why they need it and share it through a proper channel. This makes data sharing more secure, so citizen information doesn’t fall into the wrong hands.
This also means agencies can request data from others rather than collecting the same information from citizens twice. There would be no duplicate datasets, so there would be less risk.
Governments need accurate and complete data to design better policies and address citizen needs. A centralised portal can help them identify gaps and extract value from their data more easily.