Decision Support Systems and Business Intelligence
During my recent university studies (2021-24), I had a subject called Decision Support Systems and Business Intelligence. This included theoretical background on Decision Support Systems (DSS) and Business Intelligence (BI) as key tools for business decision making. DSS is a set of systems that use data, models and analytical techniques to help managers make informed decisions. These systems allow complex problems to be solved quickly and efficiently, especially in such situations where human intuition alone is not sufficient or gathering information without the complex information system would take more time. So the advantage of using such systems is not only being more precise in decision making, but making the required decisions on the right time.
Business intelligence, is a broader concept that includes data collection, processing, analysis and reporting, so all the preparations before decision. BI tools and systems are designed to provide visibility into company processes, or if external (cooperative or competitive businesses) data is available then to identify market trends and support strategic decision-making. Business intelligence enables companies to convert their data to a competitive advantage. Both DSS and BI systems share the common goal of providing the right information to decision-makers at the right time, thus increasing the efficiency and success of the organization.
On the practical side of the subject we were taught the use of MS Excel, Power Query, and majorly MS Power BI. ETL and analysis processes were practiced both in Excel and Power BI.
The exam involved theoretical and practical questions to answer, and problems to solve, as well.
Power BI is considered the most powerful tool from the above mentioned list.
The problem solving task was about a dataset of a Flower shop, selling different types of flowers to specific people (Customers). One of the questions was to demonstrate the income of the shop according to the different flower types.
It seems to be evident that the light blue colored product (Flower A) had the highest impact on the revenue.
After the exam, I went further in analysis and tried to further explore the data and searched for preferred flower-buying-day of the Customers. I would bet for Friday if I count on casual (romantic) Customer's surprising their beloved ones before the weekend starts..., here you are:
Friday is the busiest day of the Flower shop.
I was right. Maybe the justification is not the real one, but that could be verified only with further investigation, on the field, supplementing the dataset with additional information.Here comes a slight inaccuracy of the BI software to point out, in the legend the days are not listed in the order of the day number in the week, but in spite of this discrepancy the colors and day names can be paired.
This complex figure demonstrates that however Flower A was not bought in the largest amount but taking into account its highest price (3000 unit) it has the highest impact on the income, as it was also demonstrated in a simple way on the first bar graph above.
Day numbers should be matched with their English names for which I generated (typed-in) a new Table and connected to the appropriate Day data extracted from Invoice date (datetime) data column.
Another aim of the story exploration was whether there are people who insist on buying one type of flower or always the same type of bouquet. Since the Customers and Flower types had names (alphabet letters) these I converted to numbers with specific function, creating new columns and then I plotted the data:
The (scatter) plot involves Flower types (1-6, except 4) and Customer ID (1-13), bought Total quantities of the specific flower types and Unit price of the specific flower (see in the color legend on the top-left corner). Flower 1 and 2 were bought throughout the most Customers (the most spots present along the horizontal axis). Customer 12 bought the largest amount from one type of flower: Flower type 3 (the most regarding all the flower types), as it can be seen from the largest spot size on the plot, which reflects total purchased quantities (of specified Customer). Whether it was bought once-in-a-time resulting in an enormous bouquet or bought the same flower plenty times, it cannot be defined from this plot.
The spots are colored according to unit price which are in increasing order from top-to-bottom, as you can see from the Unit price legend on the top-left corner of the image.
To make it easier to compare the impact of each flower to the income I glued an additional horizontal bar graph to the plot, precisely in the same order as the flower types appear on the plot:
Note: Power BI pops up a legend with defined data related to the data-spot in the background if you hover over the spots. It is very helpful for beginner presenters who stick to the data and want to emphasize their explanation with the right numbers. This functionality is not demonstrated here.
No comments:
Post a Comment