Laundry: What’s AI got to do with it? 

Step 2: Sort according to washing instructions:

You can’t wash wool on a hot cycle.  

It’s important to understand the detail about the data you’re gathering – what are the different fields used for, which system is the most trustworthy source for the most up to date information about a particular information point, are there constraints about the purpose for which the data can be used?  For example, you might choose to trust the address provided by your housing system more than for example your social care system.  You may record information about someone having died in different ways across each source system – or it may have been recorded in one but not in all.  We need to validate the data – to organise it - so that we can handle it appropriately.  

The woollens have been separated from the school uniform and the sports gear.   

Step 3: Sort by colour and fabric requirements & wash 

Ever sent your kid to school with pink or chewing gum-coloured socks – or worse still, shrunk a new woollen jumper? I must admit to several failures on this over the years…  

Next begins the all-important process of matching the data to establish which records pertain to an individual or an entity like a house, household or an asset.  As I have been learning, it’s a bit more complicated than sorting the washing.  It involves being certain that you’re definitely matching records which pertain to the same person/asset – accounting for typos, English “translations” of foreign names and the like.  At Sentinel Partners we have some very clever tech to do this swiftly and to reduce the number of records which need to be checked by real people – in face our data matching is unparalleled, yielding unbeatable results. Each organisation has a different tolerance level about the amount of records which are quarantined for checking.

Some are happier with a hotter wash – others have a larger pile of hand washing to do.  Add detergent and watch the magic happen.  

Step 4: Sort your clean washing by owner and purpose 

Back to the clothes basket; sorting by purpose, origin, and owner is what we all do when doing the laundry. A smart business shirt might well be cotton with the same washing instructions as a cotton rugby shirt – but we know from experience that they are different garments with different uses. The smart shirt needs a good iron and hanging up in the wardrobe, the rugby shirt needs to go straight back in the sports bag for the next game.  

Data also needs sorting: knowing that despite many attributes being similar, the history and purpose of the records varies greatly. The basics like forename, or address detail, can mean one thing from one place and another from another place: the difference between the name you go by and what is on your passport, the difference between your registered address, and where you live... Sorting data by cohort, and sorting with context of information in mind, is the difference between playing rugby in a business shirt or a social worker attending the wrong address.

By using the latest matching and ML technologies you can now begin to sort your data by cohort and other vectors/categories – groups with similar issues, geographies, deprivation indices, long term conditions… the list goes on depending on where you’ve been able to get the information from. This is the foundation which allows you to shape place services based on insights drawn from many different sources – to, for example, develop improved early help approaches, to predict falls, or to predict which population health approaches could have the biggest impact.