Non-Brand Data Podcast

Machine Learning models can be untrustworthy for business people as the result often isn’t understood enough by them. We can have the best model, but it would be useless if businesses did not use it. The best way to relate to companies is to explain why our model provides certain output. In our previous NBD Lite series, we discussed techniques to interpret our machine learning model with PDPs, whi...

Model evaluation is a must-do activity when developing the machine learning model. In the case of the classification model, we have metrics more than accuracy. There are selections such as Precision, Recall, and F1-Score, depending on where we want to focus. The F1 score is usually the go-to for balanced precision and recall metrics. However, do you know that we could tweak the score even further...

One of the most tedious tasks of data scientists is to loop through the data. It’s a task that can be daunting yet necessary. Luckily, there is the Python’s built-in enumerate that could help the looping process. The enumerate function is essential if we want through iterable objects while keeping track of the index. The function would be cleaner and more readable without additional dependencies....

The regression model is popular and has been standard in many businesses, especially in the financial industry. It’s become the “standard” because of how easy it is to develop and how interpretable they are. The interpretability comes from the Regression Coefficient. However, there is another concept called Confidence Interval (CI) that you acquire when developing the model. So, why is CI importan...

Missing Data is a data point. If you drop them as it is, you might miss so much information from the missing data. Sometimes, missing data happens because of an accident or pure chance, but this is often untrue. Missing data might uncover insight that we never knew previously. In this case, we can use the missingno , a package specifically developed to visualize your missing data. How to use them?...

Hey guys! The next batch of NBD Lite series has already reached number 124 , and here is the podcast summary. The NBD Lite series is a short article series you can read in 3 minutes to improve your knowledge. Using NotebookLM, I produce an  AI-generated podcast  based on the Lite series for #17 to #24, hosted by two people who talk to each other. If you like listening instead of reading, this podc...

An imbalanced dataset is always a problem in tabular classification, but it naturally occurs. There are many Imbalance cases, such as Fraud cases, cancer patients, spam emails, etc., to name a few. In truth, Finding a pure, balanced dataset in real-world cases is harder. So, we need to be content with the imbalances that occur. However, higher imbalance intensity would skew the machine learning mo...

Hey guys! The next batch of NBD Lite series has already reached the number 16 , and here is the podcast summary. The NBD Lite series is a short article series you can read in 3 minutes to improve your knowledge. As I mentioned in the previous post, I want to experiment with new things, so I've created a  podcast summarizing my Lite summary for the week . So, it’s an AI-generated podcast based on t...

Hey guys! You might or might not know, but recently, I tried to launch my NBD Lite series . It’s a short article series you can read in 3 minutes to improve your knowledge. I want to experiment with new things again, so I've created a  podcast that summarizes my Lite summary for the week . So, it’s an AI-generated podcast based on the Lite series within the week and hosted by two people who talk w...