Dataset acquisition issue

[camillakathy]

Is anyone having issues in fetching the California dataset on Jupyter?

[shahriar-rahman]

Yes, encountered the same problem a few days ago.

[FarhanTanveir]

Try upgrading the dependencies and check whether its working.

pip install scikit-learn --upgrade

[camillakathy]

I did, but it's still not working. 😕

[shahriar-rahman]

I've upgraded the pip and scikit with Python 3.10, but the issue persists. Is it working on your end?
Found some discussions that will be able to shed some answers. One of which had the same traceback errors:
scikit-learn/scikit-learn#5051
scikit-learn/scikit-learn#5953

[camillakathy]

Okay, I'll check out the linked discussions. Hopefully, this could resolve the issue.

[FarhanTanveir]

I checked the traceback issues they were facing, and tried an alternate approach but couldn't get the download to work yet.
All other datasets can be fetched and work perfectly fine, it is the only one causing issues. We might have to download it manually.

[josh-scottt]

Does any know how to download it manually?

[shahriar-rahman]

I have an idea. The process works with IDE like PyCharm. So you can download and store it as a CSV extension. Then load the dataset in Jupyter and proceed with the EDA. Here's the code:

import pandas as pd
from sklearn.datasets import fetch_california_housing

# Acquire the data
california_data = fetch_california_housing()

# Construct the dataset
features = pd.DataFrame(california_data['data'], columns=california_data['feature_names'])
label = pd.DataFrame(california_data['target'], columns=california_data['target_names'])
dataset = pd.concat([features, label], axis=1)

# Store the dataset as a CSV extension
output_path = 'california_housing_dataset.csv'
try:
    dataset.to_csv(output_path, sep=',', index=False)
except Exception as exc:
    print(f"An error occurred: {exc}")
else:
    print("Conversion successful.")

[camillakathy]

It works now. Not sure why it didn't work on Jupyter but at least the data fetching issue is solved.

[josh-scottt]

@shahriar-rahman, did you complete it?

[shahriar-rahman]

Almost done, I pushed all the codes here.

[josh-scottt]

Cool! I'll check it out.