The Home Office’s data team rolling out a set of data analytics tools, developed without extra investment, to the business over the next six months.
The analytics capability includes internally developed algorithms and merged datasets and databases to help business users spot anomalies and monitor crime and fraud reporting.
This tool could plug a funding gap now that the government department responsible for immigration, counter-terrorism, police and drugs policy will have its budget halved over the next five years, according to data architecture and analytics lead John Holland.
"Home Office loses fifty percent of its buidget over the next five-to-six years...the scale of cuts means the business has no choice but to use data and analytics to make decisions more quickly. We have six-to-nine months to show that we can take models out of the data lab, or the science realm, and embed them into the business model to significantly improve how the business is run”, Holland said during Cloud World Forum in London yesterday.
Using algorithms and the expertise of its data scientists, the Home Office has already spotted outliers including poor data entry for crime reporting - which has lead to better education around reporting.
A report published last year found that a lack of good quality data prevented the department from managing its immigration case workload and hindered accountability.
A sample check on the casework information database found that for 84 percent of cases where people were removed from the UK, the department did not hold the minimum necessary information such as their address or postcode, before their removal.
The role of data scientists
The role of data scientists within enterprise will be “critical in the future“ due to their ability to “think where the data is and how to construct models to apply algorithms against - as well as engaging with a customer to find insight for them”, Holland added.
When the Home Office’s data modelling began, “from scratch”, it used insight from data scientists that have been trying to do the job manually for years” as a starting point to figure out where to get the most valuable data.
Keep the data dirty
While data cleansing has always been a top priority for data warehousing, Holland’s team found insights in anomalies in their raw data have helped detect fraud, for example.
“The cost of storage is cheap. Keep all that raw data...You want to be able to go back to the initial data to identify repeated behaviours - the sorts of things people could be doing deliberately, that machines tend to correct”, he advised.
Agile in government
Tools based on the analyst’s data models will be rolled out to Home Office employees and customers in “six month sprints”.
“The Home Office is 220 years and it is still running the same processes. The introduction of Agile thinking into Government development in last five years has radically changed the way things are done. Our analytics development has really embraced that.”
Holland said this release cycle was, “way, way, faster than Government has ever moved in the past.”