Last week we heard about how George’s daughter, Rita, obtained better education and career advice than her cousin because she lived in Bigdataland, rather than the UK. This week George and his wife have to recover from the shock of a break in at their house, but it is made better for them by the support that they receive in Bigdataland…
They say that coincidences are not accidental. However sometimes they can still amaze. George and his wife had just come back from a visit to George’s brother, Gerald, who lived in the UK. Whilst they had been there a significant part of the conversation had been around the burglary at Gerald’s house. Gerald was not at all happy. “All I got was an insurance number from the police. They didn’t seem at all confident that they would be able to solve the case. What really upset me was not the TV or stereo, which can be replaced. It was the personal things, like the locket with my daughter Rachel’s first hair”.
George recalled this conversation with his wife, Susan, as they drove home. “I don’t know what I would do if the same thing happened to us” she said. Sadly she was soon to find out. As they got home, they found that their back door window had been smashed, someone had gained entry to their house, and stolen a large proportion of their possessions – and their locket with Rita’s first hair was amongst the items that were gone.
The police came soon after the call was made. “I suppose you will just give me a crime number” George sadly intoned. “Not at all Sir” replied the officer. “There is a 90% chance that we will catch the culprit and an 80% chance you will retrieve your possessions”. George was interested – how could this be done. “Well”, the officer told him. “Firstly using big data meant that we were able to target all of our resources much more effectively. This means that we can put more resources on things like burglaries that would not otherwise get attention. Big data also gives us a number of tools to help us catch the culprits quickly as well. We can use accurate profiling to identify a very short list of potential burglars. We can also use the description of goods that you provide to compare with goods being sold either in second hand shops and car-boot sales, or on-line. Once we find the sellers, we can quickly find the culprits, and the rest of the goods. Finally we pool the data from a wide range of databases to link to any interesting details that we find here, and this again leads to a short list of suspects.”
The officer was as good as his word. Within two weeks, most of their possessions had been returned, and the police informed them that the culprit had been apprehended, and had already confessed. Best of all, Susan was able again to hold the locket with Rita’s hair that meant so much to her.
Sounds a bit too hopeful? Our story is certainly feasible and requires none of Phillip K. Dick’s science fiction from Minority Report where the Department of Justice predicts crimes and arrests people before they can commit them. A recent report shows that in Manchester around 15% of burglaries are solved, so there is certainly room for improvement.
For years there have been stories about individuals tracking down their own stolen goods on eBay, and today LeadsOnline aggregates data from marketplaces like eBay and pawn shops and makes it available to law enforcement. Anything one man with a search engine can do, Data Science can do bigger. A database of stolen goods and another of goods for sale can be mined to identify trends and flag up high risk sellers/listings for further investigation.
Today’s modern police force can allocate their resources more effectively, freeing up capacity for lower priority work like burglaries that, while non-violent, are still important to individuals. The Santa Cruz, California Police Department uses big data to identify crime hotspots and target them with police officers. This allows for more officers where needed, less where not, better outcomes, and more capacity overall. Even concerned citizens can access crime maps in order to assess their own risks.
States have been maintaining lists of known criminals for some time, and here in the UK they will certainly search for any fingerprints found at a crime scene they are investigating. However, the FBI is taking things to the next level with a Next Generation Identification system that includes facial recognition.
Other datasets are increasingly available to law enforcement. Phones have become “the virtual biographers of our daily activities” as an American police manual supposedly states. Extensive location data and communications logs are available. Banks can be on the watch for unearned income deposited to accounts. Payment card transactions also leave a detectable trail. There are obvious concerns here, but as that dialogue plays out and our data privacy laws mature, law enforcement will be able to use this data in well defined and effective ways.
And finally, a police force this analytically enabled will know, track, and set goals regarding their outcome statistics. A victim of theft wants to know the likelihood of getting their possessions back, just as a victim of fate wants to hear their prognosis from their doctor.
So this week we learned what can be done in Bigdataland to fight crime without anything overly Orwellian, although if a criminal brags too much on Facebook, they might get caught. Join us next week as we venture back to Bigdataland to find out more about the benefits of living in a society where data is king…