Ask the right questions

Whether you’re embarking on a smaller data analytics project or one on a larger scale, you’ll need to ensure you’re asking the right questions and thinking with a data analytics mindset. This is where we need the constant re-evaluation and the ability to pivot that we’ve mentioned, because data analytics projects are dynamic by nature. Once the data analytics proposal has been approved and the project has begun, new information and context is added, so the scope and direction of the project might change. And this change in scope and direction can change the way you interpret the numbers and outcomes.

As we already mentioned, the level of scrutiny and formality of the process will vary depending on the organization and the data analytics project. Whether this evaluation happens as a conversation, internal thought process, or a more formal document, it is important that the people tasked with evaluating or deciding on the merits of a data analytics project still follow a similar thought process and approach. Approaches may differ between organizations and analytics teams; however, the basis of a good data analytics project remains similar: having an open mind. In this video, Shruti Ahuja takes us through her iterative evaluation process which she has developed over her years as a data analyst.

This is not to suggest data analytics projects don’t have any formal assessment processes or criteria. The fact these projects are iterative and are open to hypothesis and testing means that we need to ensure we’re keeping the focus on actual business requirements and understand the organization’s capabilities to resource shifting project requirements and scope.

Let’s look at some questions that can be answered by the business and data teams respectively that should help with the ability to keep projects on track and our evaluation of suitability and resource allocation.

Questions for the business teams

While not a complete list, here are some of the questions the business-focused teams might ask when evaluating a proposal for a data analytics project:

Proposal:

• Does this proposal fit into our strategic and operational goals for the organization?
• How important do we feel this proposal is going to be with regard to prioritization and resourcing?
• What resources are we committing to the project?

Objectives:

• What are the business aligned goals and objectives attached to this project?
• What are we wanting to find out and why?
• Is what we are setting out to explore viable?

Project:

• Is the proposed project well defined?
• Who are the stakeholders that need to be consulted and included in this project?
• What is the value of this project to the organization?
• What are the risks?

Expectations:

• Does everyone have clear expectations about what can be achieved and what can’t?
• Are we able to project the likely success of the project beforehand?

Results:

• Have success metrics and objectives been clearly defined?
• How will the success or outcome be measured?
• Who is responsible for measuring and assessing the results of the project?

Questions for the data teams

We’ve already looked briefly at some of these when it comes to gathering relevant data-sets but here are some more examples of questions (specifically on the data) that we might ask to evaluate the proposal:

Data strategy:

• What variables and data-sets do we require to begin this project and carry it through?
• Will the proposed data-sets, data analytics stages, and identified variables get us to our goals?
• Do we need to evaluate and adjust the data-sets and variables further?

Data requirements:

• Is any of our current relevant data already in the correct format or does it need to be transformed?
• If we don’t have all the data-sets required, are we able to collect the missing data-sets internally?
• Are the data-sets we require accessible?

Data collection:

• Do we have the time and the budget to collect the data we need?
• If not, how much will the data collection cost and how long will it take?

External data contribution:

• Is it more cost and time efficient to bring in those missing data-sets from external sources?
• How long will it take and how much will it cost?
• What data-sets can only come from external sources?
• Can we get access to them?
• Do we have to transform them?
• Do we have the budget for that?

Data security and privacy:

• Do we have permission to use these data-sets in the way we intend to?
• Do we have the correct data security protocols in place?

If any of these questions can’t be answered or return an outcome that makes the project not viable, we might want to ask some further questions. Some of these might include:

Strategy change:

• If this investigation isn’t viable, how else can we reach our goal?
• Is there a simpler, more time- or cost-effective solution?

Objective change:

• Can we still reach our objective without these particular data-sets?

Infrastructure change:

• If this investigation is of high strategic importance, how can we better resource our data analytics infrastructure and processes to enable us to carry out these types of investigations?

Outcome change:

• If not, can we adjust our focus and projected outcome to ensure a meaningful investigation?