The Secret Of Good Reports, For Beginners

While creating a BI report, it is of the utmost importance that the underlying database is of a good quality. The underlying database defines the quality of the information that you are going to get from the report. After all, if the foundation of the building is not strong, how can we expect a strong building that can endure hardship?

There are two approaches to creating quality reports. The first approach is to understand the structure of the database and trying to figure out what type of information can be extracted from it. This approach works well for the type of databases that are already in use. Changing the structure will not make sense here. Thus, you may need to put in some extra effort to modify or manipulate the data to get the desired result.

The second approach is to design the database first, with the end result in mind. It is not just about storing the data, it is also about getting the most result out of the data, that also, with ease.

For this example, the data source will be a table in an Excel sheet. However, this is true for any tabular database.

The common mistake that beginners make is that they treat a table the same way they treat an Excel sheet. Yes, it is true that they look similar and similar things can also be done. But the most important distinction that you should always keep in mind is that there are certain rules you need to follow in order to create a table in an Excel sheet. While if it is just a spreadsheet, you can format it in any way you see fit.

Here is a simple example.

Let us compare the values of 3 product ratings of three shoe brands over few years.

First approach

Year 2004/05 2006/07 2008/09 2010/11 2012/13 2014/15
Product 1 0.83 0.43 0.32 0.45 0.5 0.42
Product 2 0.66 0.33 0.43 0.023 0.43 0.55
Product 3 0.35 0.35 0.35 1.35 1.35 1.35

For bare human eyes, this data looks easy to comprehend.

Now, we create a clustered column chart for the given data. It will look something like this,

Here, we have products at x-axis and values over years on the y-axis.

Second Approach

The data set for the second approach looks like this.

Year Product 1 Product 2 Product 3
2004/05 0.83 0.66 0.35
2006/07 0.43 0.33 0.35
2008/09 0.32 0.43 0.35
2010/11 0.45 0.023 1.35
2012/13 0.5 0.43 1.35
2014/15 0.42 0.55 1.35

Now, as for creating the same chart, the chart will look like this.

Here, we have years on the x-axis and product value on the y-axis.

Now, when we compare these visuals, we can easily understand the data from the second approach is better than that of first one. This has happened because of the underlying data structure.

Though the data is easy to understand in the tabular format, the first approach is a wrong approach because the column names and column values do not follow the principles of designing a database. But, for the second approach, the table has been properly structured, hence giving better visuals and quality report. With this type of table structure, querying on to the database will also be easy and efficient as compared to the first one.

Thus, spending some time on the designing of the database will definitely be a good pay off for creating quality reports.

Leave a Reply

Your email address will not be published. Required fields are marked *