5 Tips for Your Next Data Analysis Project

In the business world, we create an overwhelming amount of data every day. This data can be extremely useful to us if we know how to convert it into meaningful information. Once you have meaningful information, it can be interpreted, which will allow you to draw conclusions and take actions.

I have often wished that there was an all-knowing device that answers any question I asked it. Search engines like Google, Bing and Yahoo are getting closer, but there are still two problems: 1) if I don’t know what to type in the search box, then I will never find my answer, and 2) I may know exactly what to type, but it will never be found because there is a ton of data that isn’t available. This all-knowing device may not exist today, but when it does, I hope they call it the Knows All Things or KAT for short. By the way, Apple hasn’t even figured it out yet because I argue with Siri all the time.

Recently, I have experienced an increase in business owners and executives sharing with me their frustration about their inability to gather and analyze their organization’s data. They would love to have a KAT. I have others who don’t even realize the value hidden within the data they are gathering.

The point to all this is that data is an important part of our lives and that isn’t going to change anytime soon. In business, it is even more important to figure out how to gather and interpret it. If you do it today, it could give you a competitive advantage. If you don’t do it tomorrow, you may go out of business.

If you have been waiting, it is time to refine your data analysis skills. Below are five tips to use during your next data analysis project.

1. Obtain as much data as possible. When you start obtaining data, a natural instinct is to gather exactly what you will need to perform your analysis. Then at some point during your analysis, you wish you had more data – it could be for another time period or additional data fields. This can create challenges and inefficiencies down the road depending on the complexity of the data analytics performed. Having the additional data may slow down processing speed slightly or create some “noise” when evaluating results, but these two things are nothing compared to the headaches and rework that could happen if you have to go back and add data, then analyze again. Remember, the excess data can always be removed very easily at the end.
2. Understand the data. Knowing the data formats, fields, and types is imperative to planning and executing any data analytics. This seems intuitive, but sometimes things are overlooked or assumptions made that can impact your analysis.
   • Data is formatted in different ways and companies can be saving similar data in two places in different formats. As an example, the date I met my wife, March 23, 1998, can be formatted in many different ways. 3/23/1998, 3/23/98, or 23/3/98 are just three examples.
   • The data fields relate to the available data for a record. If you are using Excel, it would be the columns in the spreadsheet. We usually understand most of the columns in a data set, but in almost every data analysis project, there is some unknown column hanging out there.
   • Data types would be the actual data included in a data field for a specific record. For example: the data field could be “UserID,” and the data type could be “JHS.” If you don’t get a data table that provides a key to translate “JHS” to “John H. Smith,” you may not be able to fully interpret the results of your analysis. Another example could be a data field that contains order status, and the different data types are “CAN,” “OPN,” “CMP,” or “HLD.” You probably should gain an understanding of what these data types mean.
3. Select the right tool. There are several tools available to use and picking the right tool or tools for the job is very important for efficiency. I have seen many times in my career when people use the same tool for every data analysis project because they are comfortable using that one tool. This is a huge mistake. Think about a painter or a mechanic and all the different types of brushes or tools they have to do their jobs. For most data analytic projects, you can get by with knowing how to use Excel, Access, and IDEA or ACL. If you are working with extremely large data sets, you may need to use SQL. I will come back with more details in a future blog.
4. Perform analysis. Again, this topic will require another blog, but I will try to hit some high points related to performing the data analysis. It is best if you set some goals upfront. What are your goals? Are you looking for certain trends, or who may be involved with certain types of transactions? I like to run a few preliminary analytics just to gain an understanding of the data. As an example, I will summarize the data on certain key fields (i.e., customer name) or determine the first and last transaction dates or determine if any transaction amounts are missing. Finally, you have to develop a plan and steps to execute your analysis.
5. Summarize and report. There are a lot of things to consider when summarizing and reporting your findings. You should make sure you know your audience. You may have performed the greatest data analysis in the history of the world, but if you screw up presenting it, your audience will discount everything. I believe this could be another topic to dive into a little deeper on another day.

I hope you enjoyed this blog as much as I loved writing it. As you can tell, it is a little longer than most, and there is so much more I need to tell you about data analysis. I have a feeling there will be more blogs like this one in the future. 

For weekly insights on fighting financial fraud, click here: