Decision Mechanics

Insight. Applied.

A passion for decision analysis

By

When discussing the value we offer to clients, the focus is often on technical aspects—experience, understanding the psychology, skill with tools, processes, etc.

However, it occurs to me that often much of the value we bring to the table is energy and a fresh set of eyes. In many circumstances decision-makers have decision fatigue. It might be due to the number of decisions they have to make, or how long they have spent mulling over this particular decision.

When you are exhausted you make poor decisions.

Engaging the support of a decision scientist gets you someone who isn’t drained. In fact, it gets you someone who is excited about looking at your decision in detail, and from all angles. This work isn’t a burden to the consultant. It’s their job.

No-one likes to admit they are just plain exhausted/exasperated. But you can often see the weight of the world lifting off the shoulders of decision-makers when we come in to help.

Photo by Ian Schneider

JavaScript for data science?

By

Arquero is a data manipulation library written in JavaScript. It’s modelled on the excellent R dyplr library.

JavaScript isn’t known as a data science language. However, it’s probably the best language for data visualisation—and Google’s relentless optimisation of the V8 engine means it has better performance that Python or R.

The JavaScript ecosystem lacks extensive data science libraries, but Arquero coupled with, say, D3.js makes for a capable data visualisation platform.

Photo by Anton Darius

R now has native pipes

By

The latest release of R introduces an F#-style pipe operator.

Pipes have been available in R for a while, via the magrittr package. However, with the release of R 4.1.0, pipes are supported in base R.

The syntax has changed from that used in magrittr. The magrittr pipe operator is %>%. The new base R pipe operator is |>.

So, instead of

df %>% head()

we have

df |> head()

R 4.1.0 also introduces a lightweight, lambda-style function syntax, which can be usefully combined with the new pipe operator when piping to an argument other than the first one.

iris |> {\(x) lm(Sepal.Width ~ Sepal.Length, data = x)}()

Data cascades—the impact of data mismanagement

By

mount of garbage

It’s data science, folks. It lives and dies by the quality of the data.

Google Research recently published a paper where they argue that machine learning solutions are being undermined by a lack of focus on data quality issues. They note that

[…] data is the most under-valued and de-glamorised aspect of AI

and that data is

[…] viewed as ‘operational’ relative to the lionized work of building novel models and algorithms.

Ironically, data science arose from statisticians’ disinterest in the collection and wrangling of data. Revisiting the sins of the father, I guess.

The Google researchers point to the prevalence of data cascades—upstream events that have compounding negative effects on project outcomes.

92% of AI researchers interviewed for the study had suffered from a data cascade.

Four categories of data cascade were identified.

  • Interacting with physical world brittleness
  • Inadequate application-domain expertise
  • Conflicting reward systems
  • Poor cross-organisational documentation

All of these issues conspire to rock the very foundations of the models we increasingly rely on.

Data quality is hard to get right. It’s a much harder problem than model development. And, while the specific choice of model is often unimportant, the same is never true for the data that is fed into it.

One reason data quality to so hard to achieve and maintain is that it’s a process problem—often involving multiple organisations and stakeholders.

As the authors of the study lament,

Data quality carries an elevated significance in high-stakes AI due to its heightened downstream impact, impacting predictions like cancer detection, wildlife poaching, and loan allocations.

We need to stop fetishising algorithms at the expense of data. Tutorials on machine learning libraries and Python are smeared across the Internet. We need to promote and reward good data hygiene.

The consequences of continuing to undervalue data work are stark.

Garbage in, garbage out.

Photo by Antoine GIRET on Unsplash