Yinka Akangbe For gradient boosting, if you're familiar with python you can look at this example: https://scikit-learn.org/stable/auto_examples/ensemble/plot_gradient_boosting_regression.html#sphx-glr-auto-examples-ensemble-plot-gradient-boosting-regression-py
CNNs are mostly used for image data. I don't think using a CNN for a purely numerical regression problem like yours is very sensible.
I downloaded the dataset and went through the features and proceeded to do some preprocessing after which I built a pipeline to train a XGB Classifier and a Random Forest classifier models over it.
The XGB (with hyperparam tuning)has a MAE of 0.062 while the Random Forest has a MAE of 0.050.
Like Christian Schmidt mentioned, using a CNN for this task is an overkill but you can use a simple feed forward dense-net for this task.