Human Centred Data Science
We build original datasets that can detect shifts in consumer preferences and identify growth audiences for our partners.
The era of big data promised with algorithmic certainty that we would reach new depths of understanding and predictive power over human decisions. Yet time and again we see that this results in reductive logic—the echo chamber of a news feed algorithm or the GPS which only offers the most direct route, not the scenic one.
But what if, rather than unrelentingly collecting more data, we ask sharper questions of the numbers and celebrate the individuals behind them, not the most obvious commonalities?
What we search, what we watch, what we publicly share, what products and services we value—every data source reflects a different aspect of behaviour. We emphasise layering relevant datasets rather than relying on one source and scale. And we have learnt, through this approach, that the magic often comes from the contradictions between datasets.
Our goal is to detect and model shifting patterns in language and behaviour, then bring to life the people whose actions create these shifts. We go deep into the context in which patterns are created, interrogating data sources (platform mechanics, incentives, biases) and culture (norms, flux) in order to shed light on why shifts are happening and lay out the future directions of change.
We call this Human Centred Data Science.
How we do it
We are a team of engineers, anthropologists, data scientists and brand strategists.
Whenever we take on a new question, we start with a multidisciplinary team to ensure that each of these perspectives is represented.
The universe of open data is ever-expanding, so we experiment—a lot. In a spirit of constant innovation we add sources and processes that make our work better and discard those that no longer feel useful. The more that lives on our cutting room floor, the sharper and more focused the final output.
Independence from investors, ad sales and data providers gives us the freedom to provide an unbiased view. We pursue the most relevant sources from anywhere in order to build the richest possible perspective on the people we set out to understand.
Working with us
We look for partners, not clients.
We value working with people who are nimble; excited to feed into iterative development and improve what we do together rather than settling for more of the same.
Our work is experimental. We always start with a small pilot so we can test and prove value before scaling commitment and investment.
That means we focus on fewer, deeper relationships where we get time to immerse in the businesses we work with, and deeply understand their category and commercial priorities.
Join our team
There’s no ‘type’ of person who works at Synthesis, and that’s kind of the point.
We’re a team who come from all sorts of backgrounds, from computer science, digital anthropology, engineering, advertising, linguistics and law.
We are fiercely protective of our positive, supportive team energy, collaborative goal setting and spirit of Plussing. Together we experiment, challenge ourselves, build and rebuild, innovate and improve our thinking, processes and ways of working.
Apply if you are looking for a small, startup feel, you are excited by smart ideas and are a self-starter who is comfortable with ambiguity.
Vacancies are advertised on JobsBank and LinkedIn, and we welcome speculative applications. Please send your CV and a cover email stating your interest in Human Centred Data Science to work@synthesis.partners
Learn with XDS
We believe in the power of Human Centred Data Science. It’s how we do what we do.
We want to open up possibilities for others via our experimental data science community, XDS, and through our academic and corporate partners.
XDS brings together people interested to creatively apply their data skills to open sources and real world challenges. We run regular events in Singapore and Amsterdam. Find out more here.
If you or your organisation is interested in learning more about the practice of Human Centred Data Science, or how to apply it to understanding change or reaching new audiences, please be in touch.