Digital analysis can help reduce lead-poisoning threats

Hey there, time traveller!
This article was published 17/01/2020 (2269 days ago), so information in it may no longer be current.

Winnipeg’s drinking water hit headlines last month — for all the wrong reasons.

Samples of tap water from Winnipeg homes revealed that residents in seven wards across the city would have to let their taps run for at two minutes to ensure lead levels fell below national guidelines for safe drinking water. This comes hot on the heels of another damning year-long investigation by more than a 120 journalists revealed shockingly high levels of lead in drinking water supplies in multiple jurisdictions across Canada.

We should worry because lead poisoning can have a whole host of physical and mental impacts on its victims. We should care because in these instances of compromised standards of basic resources, it is often the most vulnerable populations that are impacted the most. And we should act because our confidence in drinking water is essential for a fair and sustainable society — as articulated clearly in the U.N.’s sixth global Sustainable Development Goal.

Right now, we need to be asking ourselves how Canadian towns and cities can still be delivering water to citizens in service lines that are laced with lead. The problem is complex, ranging from multiple latent sources of lead within our buildings and infrastructure to an awkward split in responsibility between homeowners and water providers in terms of tracking and fixing sources of lead poisoning, mostly leaching from old pipes. 

When it comes to solutions, as is often the case, we already have the data. We just need to use it better.  

In Alberta, the environment and parks department has just released an innovative new document that champions the idea of taking existing data that reveal, however obliquely, the likelihood of a street or neighbourhood being vulnerable to lead poisoning, and then using the power of data analysis to determine where to focus energies and resources. 

Let’s say, for example, a city has records on when a building was constructed and undergone assessments, the history of the permits that have been granted, and how many people can live there. On their own, those data don’t reveal much about whether a given property’s water supplies is likely to deliver lead to its inhabitants — and we should note that very few cities have access to all these data on paper, let alone in a digitized format. 

However, seizing on new means of analyzing, such as machine learning and artificial intelligence, an algorithm can predict which neighbourhoods, streets and houses are most likely to be connected by lead piping. Then they can target specific spots in the city that are most likely to be plagued by lead poisoning and focus their limited resources on remediation efforts. 

Ironically, Flint, Mich., is one of the “luckier” municipalities to have access to much of this digitized data, and is making the most of it in an exciting new project. City officials are running their data into a machine that indicates on which streets and neighbourhoods they should be focusing their very limited resources. 

And it’s working! According to Michigan Radio, “When the city stopped using the algorithm to target neighborhoods, the lead lines found in excavations fell from more than 80 per cent to 15 per cent.” 

Provision of drinking water is a basic right, but requires tackling complex questions that involve aging infrastructure, multiple layers of governing entities and limited resources. Digital technologies are being embraced globally to address those very questions, to avert another Flint or Cape Town situation. 

It’s now time for Canada — proud guardian of much of the world’s fresh water — to follow suit.

Matthew McCandless is senior director, fresh water and executive director of the IISD Experimental Lakes Area.