What is data science?
Data science is about building data solutions that solve company challenges.
This usually involves a lot of number crunching and building “models” that take real world data and convert it into something simpler for making predictions.
Once people have come up with a model to make predictions, it then gets implemented in an operational system so that predictions can be made on demand and used to tackle those business problems.
Who is needed for data science?
Data science involves what feels like a cast of thousands. It’s not just a single data scientist that makes data science work in an organisation.
Data science is valuable when it aligns to company goals – executives need to understand what data science can achieve, what’s involved at a high level, and how to fit it into budgets, priority lists and so on.
Data science is easy when there’s good data – no data means no data science, bad quality data means people trying to build models spend most of their time cleaning data. A mature data platform with high-quality data is essential. This usually means an effective group of database administrators and Business Intelligence people.
Data science needs to be practised – you need people who can do data science. This doesn’t mean you need to acquire a gaggle of people with PhDs in machine learning, but you do need people who can think about the data, understand what sorts of models solve the business challenge, and then evaluate the outcomes.
Data science needs to be implemented – models need to be in production. Outputs from the data science team need to go in the IT priority stack and systems need to be built or implemented to allow models to be operationalised.
How does all this happen?
Firstly, by acknowledging it isn’t going to happen overnight and it isn’t a silver bullet. It’s vital senior leadership understand what data science can help them with and what’s needed to see that happen.
That could mean boosting your data quality, or looking at whether any work could actually make it to production. After all, if all of IT is working on super-important projects that’ll supersede data science outputs, why do data science right now?
You might need to hire externally for someone who can perform the lions share of the work, or you might look to grow this function organically from your existing staff. You might decide to do both.
The first few projects will be slow whilst everyone involved gets to grips with what’s involved. It’s better to start small to increase the odds of success.
As time goes on, everyone starts to understand and improve. Your data science function solves business challenges more effectively and faster. They revisit and improve upon earlier work.
What can Locke Data do to help?
Steph Locke, Locke Data’s founder, has spent her recent career building up data science capabilities inside organisations. From working at finance and security startups in lead data science roles to consulting with businesses to help them through their data science teething problems, Steph has felt the pain and helped make it go away.
Locke Data offer services to help you build your internal data science capability.
The key thing we can offer you is knowledge of where to start. Our Data Science Readiness Reviews look at where you currently are in factors impacting the success of a data science initiative and give you concrete next steps.
We offer executive briefing days for senior management to get to grips with data science and how it can feed into their strategy.
We can deliver training that equips your existing analytical and Business Intelligence staff with the necessary coding and statistical skills to enable them to do data science.