OIL & GAS ARGENTINA

This dashboard is the result of a project that was submitted to participate in the "2024 National Contest of Data Visualization" organized by the University of San Andres of Argentina (UdeSA - Universidad de San Andrés). The project aims to provide information that allows an analysis of the oil & gas activity in Argentina, linking data related to the production and reserves of oil and gas, investments made by the operating companies of the concessions and fracturing activities that this type of exploitation demands.
The source of the data comes from reports published periodically by the Energy Department of Argentina government (Secretaría de Energía).

The project, in addition to working on data visualization, presents challenges in data extraction, transformation and integration, as well as in modeling and creation of industry-specific metrics.

Below is the dashboard in Power BI, and after that you can find some images that allow you to explore more about the ETL tasks using Power BI Dataflows, the dimensional data model, and some of the measures created with DAX.

Here you can see part of the ETL work with Power BI Dataflows. To perform the ETL, I tried to implement a 3-layer medallion style architecture, being layer 1 (C1) the one that corresponds to the extraction of raw data from the source; layer 2 (C2) is fed from C1 and corresponds to combinations or joins between data from the same source (for example, consolidating annual or monthly data in the same query); and finally layer 3 (C3) which is fed by C2 and is where some more specific transformations are performed and where the keys (or ids) for the relationships between fact tables and dimensions in the data model are assigned. Each layer has multiple dataflows because the model is fed with data representing multiple facts.

The resulting data model is composed of multiple fact tables. The model follows the path of a dimensional data model. The initial view of the model presents a large number of tables. For better organization and understanding create individual views (layouts) to visualize the relationship between each fact table and the respective dimension tables, define key fields in each table, and manage the visibility of key columns that do not need to be visible in the data for reporting purposes.

In order to be able to reflect good insights from the modeled data, I wrote several DAX measures representative of this industry. All measures are contained in a measure table and separated into folders according to their purpose. This good practice allows for better understanding and management, and I used calculation groups to add dynamism to the visualizations, so that some common calculations, such as time intelligence, can be recycled and reused.
To make the dashboard more usable, I inserted some field parameters that allow to easily interchange attributes or measures in the visualizations, and when assigning names to the measures I always try to use a nomenclature that allows me to identify them properly, which facilitates the documentation and management of my data dictionary.
In the creation of measures in DAX I use variables, I apply an appropriate format to the code and if necessary I use the DAX Studio tool to explore my data model and optimize measures.