Data Analytics is the process of deriving useful information from data. For a business useful means anything that helps to achieve business goals. Therefore good data analytics is goal driven and aligns closely with business objectives and is not analysis for the sake of analysis.

Data analytics projects that follow a methodology are more likely to succeed. All projects are similar in that they follow a series of steps, varying only by the length of time spent on each step. A well organised project will take the time to consider each of these stages recognising that each are important for the successful delivery of results.

The deliverables of a data analytics project are mostly in a state of change up to the modelling phase. An agile methodology is a good way of delivering projects of this sort. The agile methodology strives for rapid delivery of solutions while welcoming changing requirements putting customer satisfaction as the number one priority.

The Question

There are broadly two categories of projects. Ones that start with a series of questions that hopefully the data can answer, and those that start with data and hope that it contains hidden patterns that can be exploited to give the business some competitive advantage. In the first case you know the question, in the second, the question reveals it's self from the data.

Bear in mind that at this point you may not have fully formed the objectives of the project. As you get more into the analysis, you can expect your understanding and framing of the question to change and get more refined.

Acquiring Data

You have your goals, now for the data! The length of this task can vary considerably. The data may be immediately available in the format you want. Or the project may have to set up complex processes in order to pull the required data from multiple sources in multiple formats in multiple levels of detail. A good piece of advice would be to start with as small a data set as possible and then start building on it was the analysis gets more sophisticated.

Exploratory Analysis

You have the data, now go forth and explore! This is where the goals of the project become useful as it will guide the initial queries carried out. this stage can also help refine understanding about the questions that hopefully the data will answer.

There are many artefacts created during this stage. Any time taken to properly document and store these will pay for it's self in the future in case you would like to revisit old analysis.

Exploratory analysis is rife with false flags and red herrings but as the name suggests, It is a time to explore the data and get a better understanding of it.

Modelling

Once you find the most feasible route through analysis and decided on the end game, you can start building the formal model of the solution. It takes ideas from exploratory analysis and build it into a useable model. At this point you should have a well defined understanding of what the final goal of the project should be and the model should align closely with those goals. There are good modelling habits you can adopt that have been discussed in this blog post.

Interpreting the Model

Interpretation is how the model answers the original questions or the goals set out in the beginning. Consider yourself lucky if the answer is straight forward. In most of the cases, the answer will be subject to interpretation and perhaps have a mathematical basis. Start with your initial question, and make sure you haven't lost your way.

Communication

How to communicate the findings to the stakeholders that sponsored the project. Is this a one off piece of analysis or would the stakeholders want to track the findings in the model over a period? At what level is the analysis being done. Executives will require summary information with the ability to drill down to details.

Decision

A good data analysis methodology provides a framework to make decisions. Proper organisation means you can make the best decision with the data available. It also means that you will be able to come back at a future date and work out why a particular decision was taken.

It is the job of the data analyst to facilitate this process and help make a decision that will justify the analysis project.