yomaio.blogg.se

I first imported JSON and pandas in my notebook. I downloaded the data from postman as you did and saved it to a JSON file. data = json.load(open('json_example.json')) I did this slightly differently from you, but I think the outcome is the same. I also parsed out the data object in the payload to make it easier for this example. I used pd.json_normalize to create the data frame. I started with reference_tweets and separated this further.

This is based on the “not so good solution” from this article. As indicated in the medium article, you can do the merge right away, but I personally prefer to do it step by step to have more control. I also renamed the new df columns to it add easier to add the columns to the primary data frame. new_df = df.apply(pd.Series)ĭf = normįor the ntions object, this was a bit more complex because the file I had up to 2 mentions per Tweet. Now you can view your main data frame, and you should see the new columns you created.Įnt_df = df.apply(pd.Series)ĭf = newĭf = new_b Your file may have more mentions per Tweet. You may also want to drop any remaining columns with the data we parsed out. Hi all, I am new here and might be asking the same question as the one that has been answered 20 days ago, but if so I didn’t really get the answer…! It would be referenced_tweets and ntions.Īlso, I have passed on the feedback that data nested in this way can be difficult to work with to the product team, for us to consider as we make future improvements. I am using the academic API to get tweets between september and november 2019 in Sweden.