How to convert binary file to pandas dataframe

Question

I use Amazon Sagemaker for model training and prediction. I have a problem with the returned data with predictions. I am trying to convert prediction data to pandas dataframe format.

After the model is deployed:

from sagemaker.serializers import CSVSerializer
xgb_predictor=estimator.deploy(
    initial_instance_count=1,
    instance_type='ml.g4dn.xlarge',
    serializer=CSVSerializer()
)

I made a prediction on the test data:

predictions=xgb_predictor.predict(first_day.to_numpy())

But, returned prediction results are in a binary file. This is a sample returned predictions:

b'2.092024326324463
10.584211349487305
18.23127555847168
2.092024326324463

How can I convert prediction data to pandas dataframe?

Answer

Hi!

It looks like you have a bytes response, have a look at [this post](https://stackoverflow.com/questions/47379476/how-to-convert-bytes-data-into-a-python-pandas-dataframe), this will help you with the transform to dataframe.

Hope it helps :)

How to convert binary file to pandas dataframe

Conteúdo relevante