Power BI

Levente Tóth

Marketing Carrier On-Time Performance During COVID-19

Due to Coronavirus, many airlines are canceled their flights all around the world, let's have a look at how the cancellation changed in the US during this period from 2020 to 2021.

Flight cancellation report in Power BI

From Bureau of Transportation Statistic, the report shows the percentage of canceled flights, depending on which Airline or State we want to analyze.

The report shows between March 2020 and May 2020 the cancellation ratio increased significantly, compared to other months.

For more findings: Click here to view the full report.

airline_pbi_with_logo

Model and Connections

The report has fact and dimension tables as well. Fact tables store the observations which in our case is each individual flight. A fact table contains dimension key columns that relate to dimension tables, for example, AirlineID contains the airline's full name, the same as for AirportID, AreaID, and CacellationID where the cancellation reason is specified.

What is star schema? Star schema is a mature modeling approach widely adopted by relational data warehouses. It requires modelers to classify their model tables as either dimension or fact.

Underlying the report has the following data model and characteristics:

airline_model_pbi

Generally, dimension tables contain a relatively small number of rows, fact tables on the other hand, can contain a very large number of rows and continue to grow over time. If we didn't use a scheme like this, the overall size of our data would be extremely huge, and it would be slow to work with.

Next read

Statistical Analysis & Machine Learning Workflow for Property-pricing
Python

Statistical Analysis & Machine Learning Workflow for Property-pricing

The goal of this project is to work on a predictive model, that generates accurate predictions for future, yet-to-be-seen data. This includes pre-processing the data, selecting an approach for identifying optimal tuning parameters, building a model, and estimating predictive performance. This approach protects from overfitting the training data and helps the model to identify truly predictive patterns that are generalizable to future data, and enabling good predictions.

Levente Tóth
Levente Tóth

© 2021-2022 Levente Tóth

All rights reserved.