3 Ways to Analyze Data

Analyzing data is such a great topic. My goal with this post is to provide you with a structured way to tackle your next data analytics project. Although there are many reasons to perform data analytics, you should always focus on the tic-tac-toe method, that is horizontal, vertical and diagonal.

I often get questions from people related to this topic, and I try to answer their specific questions. However, it is sometimes a challenge because there are so many important variables to consider when performing data analytics. If you have specific questions, please reach out to me and I’ll be happy to discuss them with you.

Before you get too far into your analysis project, set some goals and write them down. Let’s say you have identified a fraud scheme and are trying to identify when it started and who may be involved. With this example, you can set some specific goals, which you should be able to reach fairly quickly. In another example, let’s suppose that you have no clue whether fraud is occurring, but you want to explore this possibility. This one is a little more challenging. You should set specific time limits when you set your goals for this one, because without limits, you could analyze for days and never find anything.

Based on your goals, you should develop steps or tasks to try to reach your goals and answer your questions. Remember flexibility is an important part of a fraud investigation. You may need to obtain more data, change your goals, or change planned tasks. The following are three ways to analyze data:

• Horizontal Analysis – This analysis involves the evaluation of changes or trends over time. You should summarize the data by month or year, and then calculate the percentage of change between the periods. Below is an example:

• Vertical Analysis – This analysis involves the evaluation of comparing customers or vendors compared to each other or expense items to revenues or profit from the same period. You could also evaluate duplicates or missing items in a data set (e.g., check numbers when evaluating cash disbursements or invoice numbers when evaluating vendor payments). Below is an example:

• Diagonal Analysis – This one can be pretty much any other type of analysis. You can compare data from two different data sets. For example, you could compare employee addresses to vendor addresses to look for matches. The sky is the limit here and it helps to collaborate with others to determine the most effective and creative ways to evaluate your data.

Finding fraud is all about identifying anomalies, inconsistencies, or variances. You are looking for things that don’t make sense but that doesn’t mean its fraud, at least not yet. Data analytics will only take you so far. For the next step, you will need to research by seeking out supporting documents to explain the anomalies, inconsistencies, or variances. You may also interview people involved in the transactions or decision making process to determine their knowledge or involvement. I know research isn’t exactly data analysis, but it is necessary if you want to take steps to identify fraud.

Finally, if it’s there, fraud can be found. However, finding it is a combination of skill, experience, and a bit of luck. 

For weekly insights on fighting financial fraud, click here: