Official MOOC launch today!

Welcome MOOC students! :slight_smile: This forum already has quite a lot of discussion, since we used it for the in-person course (for the USF MS in Data Science program). Before posting, please search to see if your question has already been answered.

Feel free to add comments or questions to existing threads, or create new topics as appropriate. Also note the wiki threads and notes linked here:

26 Likes

Hello,

I’m trying the House Prices Competition on Kaggle, and there are several date fields in the data set.
When converting the categorical date fields into codes, is there a function to search through all the date fields and use the add_datepart function on them, instead of manually running the function on each of them?

Also, I have trouble identifying which fields I have to look for when setting the order of categories. In lesson 1, Jeremy goes straight to the UsageBand category, do you have to look at each category to see which one could be ordered differently?

1 Like

BTW, before the official launch, I have helped to edit, tidy up and standardized the header for all the wiki threads. Just minor modifications like adding lesson notes from hiromi and back and/or previous link to lesson.

Congratulations on the official launch of ML1 !
I believe there´s a minor mistake at lesson 7 title:" Lesson 7 - Introduction to Random Forests" on the website(http://www.fast.ai/2018/09/26/ml-launch/).
It would be better: “RF from scratch” or similar.
Here in BrasĂ­lia, Brazil we have already finished lesson 7 and people are loving the course. Thanks Jeremy!

2 Likes

Thanks so much - I think it’s a big improvement :slight_smile:

Congrats! Awesome this is now officially launched and we get access to the actual machine learning category! Much better than the single thread we MOOCers had to use before :wink:
Thank you very much for this course!

1 Like

@fastai1 Hi I am also trying the House Prices Competition on Kaggle and have similar doubts. Have you got any solution for this?

@rijul

You could try to loop through all columns and when the column is of dtype datetime64[ns] run the add_datepart function on it:

for col in df_raw.columns:
    if df_raw[col].dtype == 'datetime64[ns]':
        add_datepart(df_raw, col)

Tell me if this works for you.

If someone replies to one of my posts, how can I reply back to them without quoting or creating a new post?