Good to share
ONE of the projects that the UN Global Pulse has conducted so far is to compare online conversations with official national statistics to gain a better understanding of unemployment trends. This was carried out in the United States and Ireland over a period of three months.
“The main finding was that it was possible to use online data that is publicly available to provide accurate predictive indications of a pending spike in unemployment,” says I-Sah Hsieh, global manager of international development at analytics software development company, SAS Institute (SAS).
“In Ireland, there was an increased chatter about food, then housing, and then downgrading transportation that preceded a spike in unemployment. In the US, negative chatter about healthcare expenses increased, followed by negative employment chatter, and then increased use of public transportation preceded unemployment spikes.”
Since the project’s completion, he says that government agencies from various countries have shown great interest in incorporating the best practices from the project into their existing real-time reporting environments.
According to Hsieh, the United States and Ireland were chosen for the project because they were both English speaking nations, so that made it simpler to analyse data between the two countries. At the same time, however, both nations had very distinct cultures and had large amounts of official labour statistics available.
Since the success of its pioneer project, the UN Global Pulse has been researching wider applications for such real time online data feedback mechanisms.
It also carried out another project with SAS in Indonesia to track the movement of food, fuel and job indicators.
Other applications it has considered include using anonymous telecommunications call records to monitor mass migration patterns and identifying signs of financial hardship from mobile account reload patterns. For instance, when people who used to reload US$5 (RM17) each time now only reload in US$1 (RM3.30) increments.
“There are also health related projects they are considering which are similar to Google’s study on flu trends, but perhaps using Twitter or Facebook instead,” says Hsieh, adding that social media trends have proven to be better indicators than web search patterns alone.
Room for improvement
On the whole, Hsieh admits that there are still areas to be ironed out in terms of how real time online data analysis can be conducted.
“The processes used in mining data online for hidden trends and knowing which data sources are best suited for which types of problems will continue to improve with real world practice,” he says.
Nevertheless, Hsieh sees a bright future for the use of such analytics tools when placed in the hands of the UN and similarly, other national policymakers.
“Never before has the UN had the opportunity to have a conversation with all the global citizens as they draft and create new policies pertaining to complex socio-economic issues. With SAS, they now have the ability to pose a question, listen to and analyse responses from the citizens of the entire world in order to help everyone make better informed decisions,” he says.
So while it may bother us to have our data shared with the authorities, perhaps it may not always be a bad thing, since ultimately it is to our benefit if the government is able to arrive at better public policies as a result.
Telltale signs: In the united States, negative chatter about healthcare expenses led to negative employment chatter that was then followed by unemployment spikes.
Sharing data may not always be a bad thing if it leads to better public policies.
Comparing online conversations with official national statistics provides a better understanding of unemployment trends.