Why is Data like Bread?

Let me count the ways:

Data is a fundamental foodstuff for decision making?
The structure of your bread is almost as important as the raw ingredients?
People sometimes have strong opinions about the best way to fold data?
Bread occupies significant portions of conversation time within the tails.com data team?
Data carries a significant burden of.... proof?
Bread is much better when it's relatively fresh and isn't stale having been sitting at the back of the database for several weeks?!

All excellent choices, but... not the reason for this post. Writing this in the middle of Summer 2020, while much of the world is on lockdown as part of the current COVID19 pandemic, what has struck me from my conversations with data professionals from all kinds of businesses is this:

Many people are rediscovering bread, and many organisations are rediscovering data. Both require considerable time and attention but provide reassurance, grounding and direction in this bewildering time.

In the last few months we have seen public government broadcasts containing graphs mapping the spread of the virus to justify lock down measures, we have seen blogs like informationisbeautiful pumping out gorgeous visualisations to help communicate what is going and drive decision making and we've heard podcasts like More or Less covering the difficulty of forecasting in times of uncertainty. Lots of people have lots of questions, and unlike the situation we had only twelve months ago, it's very difficult for people to make decisions purely on the basis of gut feel. All of the rules and assumptions that we used to rely on to guide gut-based decision making have gone out of the window, which is leading even the most data skeptic person to start turning to the numbers for some idea of what to do.

This might all be rosy, but I think we've also seen the other side of the equation too. There are questions in the media about the reliability of numbers, that their lineage or collection hasn't been reliable or unbiased. Data professionals themselves have had to be very open about about the limits of what they can say given the available data (shown both in the podcast above), but also within tails.com day to day (and I assume other organisations) when trying to answer questions about what will happen as lock-down measures ease. The unprecedented level of open data has also empowered a cohort of citizen data scientists to delve into the trends themselves, which alone is excellent, but that also raises the question of responsibility when we start projecting future death rates based on the existing data. If someone applies poor statistics to the data (or even just a lack of context) and shouts about their results, those results can becomes news very quickly, all based on shaky ground. There is so much we don't know, and as professional investors have been telling us for years, past performance is not indicative of future results.

Personally, what fills me with hope right now is simply that the conversation is happening. Not conversation about "data" in a vague way like some of the conversations in recent memory about "Big Data" and "AI" have come across, as a complicated and mystical form of technology that will solve all our problems. Much more that we are all engaging with, and having meaningful conversations about, the real tangible concepts of good decision making which many data professionals have often struggled to convey.

We are having sensible public conversations about data quality. Digging into where our numbers come from and to what degree they can be trusted, but also not just throwing out every dataset because there is some element of residual doubt.
We are having sensible public conversations about data collection. Calls for greater and more widespread testing for COVID19 to better understand the mortality rate and the current infection rate.
We are having sensible public conversations about risk and uncertainty. Whilst we cannot know exactly how this will pan out, as a nation we are slowly getting our heads around this uncertainty and finding ways to make decisions about how to balance different outcomes, all of which are currently possible, but only one of which will actually occur.
We are seeing the reclamation of the term expert in public media. After years of blatant mistrust in experts leading to movements like the flat earth society and the coining of the term fake news, we have instead seen growing trust in public figures like Anthony Fauci and Chris Whitty.
We are having sensible public conversations about the limits of our analysis. Not all the answers we seek are in the data we could collect even if we had infinite resources, and this is leading to conversations about making decisions on the basis of principles and on logic in the absence of data.
We are all reminded that data is not the goal. It is just a tool to help us make better decisions. What matters is the decisions we make, and the actions that people on the front line take to help us reach our goals. Unless data helps us make a decisions we wouldn't otherwise make, then it has no purpose.

These are all signals that make me tremendously optimistic about the world that will emerge out of this global pandemic and the way that people will approach decision making in a post COVID19 society. And with that, I should really go and check on how the bread is rising...