How I Spent My Day | 28-06-2025

Good day, guys,
I welcome you all to my blog. It is another beautiful day of the week, and I'm very glad to share with you all how I spent my day. I am always delighted to share with you my activities for the previous day via this very diary game post, as it helps me to keep track of my activities and how I can also improve on the things I initially didn't do well for the day.

Yesterday was Saturday. Usually, I join my friends and others to play football in the field, but I couldn't do that yesterday because of the assessment I was having and struggling to work on and submit before the deadline.

Since it was a Collaboratory research project and the tweak in the topic was ARIMA, Transformer, and LSTM for 3 of us respectively, we decided to work using the same data from the OpenWeatherMap API and Kaggle. The research involves the use of primary and secondary datasets, which were obtained from the sources I have mentioned above.

Since they are not too good at using Python for analysis, after doing mine, I decided to guide them on how they can be able to use the data and analyse the 3 forecasting models in each of the works and then compare the results. So the idea was to analyse ARIMA, ML (Random Forest), and Hybrid ARIMA-ML in my case. This means that I will get results for the standalone models and then combine both to see the result, and that was what my other colleagues also did.

During the analysis, I noticed a few things with the ARIMA, Transformer, and LSTM that we used in our separate works. One of the key notable things I saw there was that ARIMA performed very poorly, but the execution time was very fast, in the same way that Transformer performed the best with a much higher processing time, and lastly, LSTM followed Transformer in terms of accuracy, but it has the longest execution time.

In the third assessment, which we will be doing, it is expected that we will have all of these comparisons made and presented in the form of a poster. So I'm already preparing for that. Also, I will be working closely with my group to see that our research is published in one of the available journals or conferences.

The goal is that everyone should have at least 3 publications on this course alone. We are working towards that and we hope to achieve it. My goal is to have at least 10 publications before I finish my Master's by Research programme, which will end by January 2026.

Guys, that was all I did for the day, and that was how my day went. It was work all through, but thank God I'm now free a bit, so I will be resting for a while as I compile the work to IEEE format for conference presentation. See you all in my next publication, my people.