VN3T Opinion: Our Understanding of Data Is About to Fundamentally Alter
What is data? A strange question. But when we think of data, we think of ledgers, and spreadsheets, but ultimately data is any information which can be observed or recorded. Data is everything. The motion of the planets, the height of a mountain, what I had for breakfast. It’s all data.
All data is useful somehow. Let’s take what I had for breakfast as our example. I ate a piece of toast, a small omlette, a tomato, and a few mushrooms cooked in a little sunflower oil. This meal represents what we call a “data point”. It can give us some indication towards some possible conclusions, but on its own it is unreliable. To illustrate, below are some conclusions that you could draw from this data point;
- There are lots of good fats in that meal, he is trying to look after his heart.
- That meal is nutritionally balanced, he is an athlete.
- That is a fairly substantial meal, he has an active lifestyle.
- There is no meat in that meal, he is a vegetarian.
But reliable conclusions can’t be drawn from a single data point because it needs to be contextualized in order to make sense. All of the above conclusions are reasonable given the data available, but in fact its incompleteness means that the conclusions are flawed;
- That meal is a heart-healthy one but there is a huge amount of data missing to contextualize it. Is this what I eat for breakfast every day? What do I eat for the other meals? Do I exercise regularly? Do I smoke? Do I drink alcohol? Am I a healthy weight?
- And 3. The same list of questions could be applied to these.
4. There is no meat in that meal. But I am not a vegetarian.
The more data that is available, the better the conclusions we can draw, and for that we need more datapoints to build better data sets. One datapoint does not only need to belong to one data set. So to return to the example of the breakfast, that datapoint could belong to any number of data sets;
- Eating habits of white males
- Eating habits of English males
- Breakfast preferences of Europeans aged between 30 and 40
- Egg consumption in North Western Europe
And so on. The point being that a datapoint is not part of one whole, but in fact can be part of an almost infinite number of cardinal data sets. There are already a ton of datasets out there that are just as focused as the examples given above, but the data contained within those is not only relevant in one contextual setting. It can be cross-purposed into other new ones.
That’s where VN3T come in. VN3T is a data market that buys and sells data peer to peer. Simple enoough, but what also happens in the background is that those datasets are cross referenced and used as a basis to create more new datasets. So for example, a buyer wants a dataset containing the number of black males in America that eat meat regularly. Let’s say that dataset doesn’t yet exist. But what does exist is a dataset showing the number of vegetarians in America which includes racial information, and a dataset showing the racial identity and gender of every American citizen. The datapoints in those two datasets can be combined to create a set which contains the data held on all non-vegetarian black males. This set can be expanded into a set showing vegetarian black people of both genders, non-vegetarian white people, American vegetarians of all races, Asian meat-eaters, and so on. Each of these datasets is distinct in its own right and saleable to the right buyer, and this is just the result of combining two datasets. The same magic can be worked by combining isolated data points, and then cross-purposing the resulting datasets to create yet more focused, targeted and granular datasets.
This novel approach will completely revolutionize the data market, and make data sale and purchase easier and more targeted than ever before.